examples of a background of the study for a research paper

Educational resources and simple solutions for your research journey

What is the Background of a Study and How to Write It (Examples Included)

Have you ever found yourself struggling to write a background of the study for your research paper? You’re not alone. While the background of a study is an essential element of a research manuscript, it’s also one of the most challenging pieces to write. This is because it requires researchers to provide context and justification for their research, highlight the significance of their study, and situate their work within the existing body of knowledge in the field.  

Despite its challenges, the background of a study is crucial for any research paper. A compelling well-written background of the study can not only promote confidence in the overall quality of your research analysis and findings, but it can also determine whether readers will be interested in knowing more about the rest of the research study.  

In this article, we’ll explore the key elements of the background of a study and provide simple guidelines on how to write one effectively. Whether you’re a seasoned researcher or a graduate student working on your first research manuscript, this post will explain how to write a background for your study that is compelling and informative.  

Table of Contents

What is the background of a study ?  

Typically placed in the beginning of your research paper, the background of a study serves to convey the central argument of your study and its significance clearly and logically to an uninformed audience. The background of a study in a research paper helps to establish the research problem or gap in knowledge that the study aims to address, sets the stage for the research question and objectives, and highlights the significance of the research. The background of a study also includes a review of relevant literature, which helps researchers understand where the research study is placed in the current body of knowledge in a specific research discipline. It includes the reason for the study, the thesis statement, and a summary of the concept or problem being examined by the researcher. At times, the background of a study can may even examine whether your research supports or contradicts the results of earlier studies or existing knowledge on the subject.  

How is the background of a study different from the introduction?  

It is common to find early career researchers getting confused between the background of a study and the introduction in a research paper. Many incorrectly consider these two vital parts of a research paper the same and use these terms interchangeably. The confusion is understandable, however, it’s important to know that the introduction and the background of the study are distinct elements and serve very different purposes.   

• The basic different between the background of a study and the introduction is kind of information that is shared with the readers . While the introduction provides an overview of the specific research topic and touches upon key parts of the research paper, the background of the study presents a detailed discussion on the existing literature in the field, identifies research gaps, and how the research being done will add to current knowledge.  
• The introduction aims to capture the reader’s attention and interest and to provide a clear and concise summary of the research project. It typically begins with a general statement of the research problem and then narrows down to the specific research question. It may also include an overview of the research design, methodology, and scope. The background of the study outlines the historical, theoretical, and empirical background that led to the research question to highlight its importance. It typically offers an overview of the research field and may include a review of the literature to highlight gaps, controversies, or limitations in the existing knowledge and to justify the need for further research.  
• Both these sections appear at the beginning of a research paper. In some cases the introduction may come before the background of the study , although in most instances the latter is integrated into the introduction itself. The length of the introduction and background of a study can differ based on the journal guidelines and the complexity of a specific research study.  

Learn to convey study relevance, integrate literature reviews, and articulate research gaps in the background section. Get your All Access Pack now!    

To put it simply, the background of the study provides context for the study by explaining how your research fills a research gap in existing knowledge in the field and how it will add to it. The introduction section explains how the research fills this gap by stating the research topic, the objectives of the research and the findings – it sets the context for the rest of the paper.   

Where is the background of a study placed in a research paper?  

T he background of a study is typically placed in the introduction section of a research paper and is positioned after the statement of the problem. Researchers should try and present the background of the study in clear logical structure by dividing it into several sections, such as introduction, literature review, and research gap. This will make it easier for the reader to understand the research problem and the motivation for the study.  

So, when should you write the background of your study ? It’s recommended that researchers write this section after they have conducted a thorough literature review and identified the research problem, research question, and objectives. This way, they can effectively situate their study within the existing body of knowledge in the field and provide a clear rationale for their research.  

Creating an effective background of a study structure  

Given that the purpose of writing the background of your study is to make readers understand the reasons for conducting the research, it is important to create an outline and basic framework to work within. This will make it easier to write the background of the study and will ensure that it is comprehensive and compelling for readers.  

While creating a background of the study structure for research papers, it is crucial to have a clear understanding of the essential elements that should be included. Make sure you incorporate the following elements in the background of the study section :   

• Present a general overview of the research topic, its significance, and main aims; this may be like establishing the “importance of the topic” in the introduction.   
• Discuss the existing level of research done on the research topic or on related topics in the field to set context for your research. Be concise and mention only the relevant part of studies, ideally in chronological order to reflect the progress being made.  
• Highlight disputes in the field as well as claims made by scientists, organizations, or key policymakers that need to be investigated. This forms the foundation of your research methodology and solidifies the aims of your study.   
• Describe if and how the methods and techniques used in the research study are different from those used in previous research on similar topics.   

By including these critical elements in the background of your study , you can provide your readers with a comprehensive understanding of your research and its context.  

How to write a background of the study in research papers ?  

Now that you know the essential elements to include, it’s time to discuss how to write the background of the study in a concise and interesting way that engages audiences. The best way to do this is to build a clear narrative around the central theme of your research so that readers can grasp the concept and identify the gaps that the study will address. While the length and detail presented in the background of a study could vary depending on the complexity and novelty of the research topic, it is imperative to avoid wordiness. For research that is interdisciplinary, mentioning how the disciplines are connected and highlighting specific aspects to be studied helps readers understand the research better.   

While there are different styles of writing the background of a study , it always helps to have a clear plan in place. Let us look at how to write a background of study for research papers.    

• Identify the research problem: Begin the background by defining the research topic, and highlighting the main issue or question that the research aims to address. The research problem should be clear, specific, and relevant to the field of study. It should be framed using simple, easy to understand language and must be meaningful to intended audiences.  
• Craft an impactful statement of the research objectives: While writing the background of the study it is critical to highlight the research objectives and specific goals that the study aims to achieve. The research objectives should be closely related to the research problem and must be aligned with the overall purpose of the study.  
• Conduct a review of available literature: When writing the background of the research , provide a summary of relevant literature in the field and related research that has been conducted around the topic. Remember to record the search terms used and keep track of articles that you read so that sources can be cited accurately. Ensure that the literature you include is sourced from credible sources.  
• Address existing controversies and assumptions: It is a good idea to acknowledge and clarify existing claims and controversies regarding the subject of your research. For example, if your research topic involves an issue that has been widely discussed due to ethical or politically considerations, it is best to address them when writing the background of the study .  
• Present the relevance of the study: It is also important to provide a justification for the research. This is where the researcher explains why the study is important and what contributions it will make to existing knowledge on the subject. Highlighting key concepts and theories and explaining terms and ideas that may feel unfamiliar to readers makes the background of the study content more impactful.  
• Proofread to eliminate errors in language, structure, and data shared: Once the first draft is done, it is a good idea to read and re-read the draft a few times to weed out possible grammatical errors or inaccuracies in the information provided. In fact, experts suggest that it is helpful to have your supervisor or peers read and edit the background of the study . Their feedback can help ensure that even inadvertent errors are not overlooked.  

Get exclusive discounts on e xpert-led editing to publication support with Researcher.Life’s All Access Pack. Get yours now!  

How to avoid mistakes in writing the background of a study  

While figuring out how to write the background of a study , it is also important to know the most common mistakes authors make so you can steer clear of these in your research paper.   

• Write the background of a study in a formal academic tone while keeping the language clear and simple. Check for the excessive use of jargon and technical terminology that could confuse your readers.   
• Avoid including unrelated concepts that could distract from the subject of research. Instead, focus your discussion around the key aspects of your study by highlighting gaps in existing literature and knowledge and the novelty and necessity of your study.   
• Provide relevant, reliable evidence to support your claims and citing sources correctly; be sure to follow a consistent referencing format and style throughout the paper.   
• Ensure that the details presented in the background of the study are captured chronologically and organized into sub-sections for easy reading and comprehension.  
• Check the journal guidelines for the recommended length for this section so that you include all the important details in a concise manner. 

By keeping these tips in mind, you can create a clear, concise, and compelling background of the study for your research paper. Take this example of a background of the study on the impact of social media on mental health.  

Social media has become a ubiquitous aspect of modern life, with people of all ages, genders, and backgrounds using platforms such as Facebook, Instagram, and Twitter to connect with others, share information, and stay updated on news and events. While social media has many potential benefits, including increased social connectivity and access to information, there is growing concern about its impact on mental health.   Research has suggested that social media use is associated with a range of negative mental health outcomes, including increased rates of anxiety, depression, and loneliness. This is thought to be due, in part, to the social comparison processes that occur on social media, whereby users compare their lives to the idealized versions of others that are presented online.   Despite these concerns, there is also evidence to suggest that social media can have positive effects on mental health. For example, social media can provide a sense of social support and community, which can be beneficial for individuals who are socially isolated or marginalized.   Given the potential benefits and risks of social media use for mental health, it is important to gain a better understanding of the mechanisms underlying these effects. This study aims to investigate the relationship between social media use and mental health outcomes, with a particular focus on the role of social comparison processes. By doing so, we hope to shed light on the potential risks and benefits of social media use for mental health, and to provide insights that can inform interventions and policies aimed at promoting healthy social media use.  

To conclude, the background of a study is a crucial component of a research manuscript and must be planned, structured, and presented in a way that attracts reader attention, compels them to read the manuscript, creates an impact on the minds of readers and sets the stage for future discussions. 

A well-written background of the study not only provides researchers with a clear direction on conducting their research, but it also enables readers to understand and appreciate the relevance of the research work being done.   

Frequently Asked Questions (FAQs) on background of the study

Q: How does the background of the study help the reader understand the research better?

The background of the study plays a crucial role in helping readers understand the research better by providing the necessary context, framing the research problem, and establishing its significance. It helps readers:

• understand the larger framework, historical development, and existing knowledge related to a research topic
• identify gaps, limitations, or unresolved issues in the existing literature or knowledge
• outline potential contributions, practical implications, or theoretical advancements that the research aims to achieve
• and learn the specific context and limitations of the research project

Q: Does the background of the study need citation?

Yes, the background of the study in a research paper should include citations to support and acknowledge the sources of information and ideas presented. When you provide information or make statements in the background section that are based on previous studies, theories, or established knowledge, it is important to cite the relevant sources. This establishes credibility, enables verification, and demonstrates the depth of literature review you’ve done.

Q: What is the difference between background of the study and problem statement?

The background of the study provides context and establishes the research’s foundation while the problem statement clearly states the problem being addressed and the research questions or objectives.

Editage All Access is a subscription-based platform that unifies the best AI tools and services designed to speed up, simplify, and streamline every step of a researcher’s journey. The Editage All Access Pack is a one-of-a-kind subscription that unlocks full access to an AI writing assistant, literature recommender, journal finder, scientific illustration tool, and exclusive discounts on professional publication services from Editage.  

Based on 22+ years of experience in academia, Editage All Access empowers researchers to put their best research forward and move closer to success. Explore our top AI Tools pack, AI Tools + Publication Services pack, or Build Your Own Plan. Find everything a researcher needs to succeed, all in one place –  Get All Access now starting at just $14 a month !    

Related Posts

What are the Best Research Funding Sources

Inductive vs. Deductive Research Approach

• Resources Home 🏠
• Try SciSpace Copilot
• Search research papers
• Add Copilot Extension
• Try AI Detector
• Try Paraphraser
• Try Citation Generator
• April Papers
• June Papers
• July Papers

How to Write an Effective Background of the Study: A Comprehensive Guide

Table of Contents

The background of the study in a research paper offers a clear context, highlighting why the research is essential and the problem it aims to address.

As a researcher, this foundational section is essential for you to chart the course of your study, Moreover, it allows readers to understand the importance and path of your research.

Whether in academic communities or to the general public, a well-articulated background aids in communicating the essence of the research effectively.

While it may seem straightforward, crafting an effective background requires a blend of clarity, precision, and relevance. Therefore, this article aims to be your guide, offering insights into:

• Understanding the concept of the background of the study.
• Learning how to craft a compelling background effectively.
• Identifying and sidestepping common pitfalls in writing the background.
• Exploring practical examples that bring the theory to life.
• Enhancing both your writing and reading of academic papers.

Keeping these compelling insights in mind, let's delve deeper into the details of the empirical background of the study, exploring its definition, distinctions, and the art of writing it effectively.

What is the background of the study?

The background of the study is placed at the beginning of a research paper. It provides the context, circumstances, and history that led to the research problem or topic being explored.

It offers readers a snapshot of the existing knowledge on the topic and the reasons that spurred your current research.

When crafting the background of your study, consider the following questions.

• What's the context of your research?
• Which previous research will you refer to?
• Are there any knowledge gaps in the existing relevant literature?
• How will you justify the need for your current research?
• Have you concisely presented the research question or problem?

In a typical research paper structure, after presenting the background, the introduction section follows. The introduction delves deeper into the specific objectives of the research and often outlines the structure or main points that the paper will cover.

Together, they create a cohesive starting point, ensuring readers are well-equipped to understand the subsequent sections of the research paper.

While the background of the study and the introduction section of the research manuscript may seem similar and sometimes even overlap, each serves a unique purpose in the research narrative.

Difference between background and introduction

A well-written background of the study and introduction are preliminary sections of a research paper and serve distinct purposes.

Here’s a detailed tabular comparison between the two of them.

What is the relevance of the background of the study?

It is necessary for you to provide your readers with the background of your research. Without this, readers may grapple with questions such as: Why was this specific research topic chosen? What led to this decision? Why is this study relevant? Is it worth their time?

Such uncertainties can deter them from fully engaging with your study, leading to the rejection of your research paper. Additionally, this can diminish its impact in the academic community, and reduce its potential for real-world application or policy influence .

To address these concerns and offer clarity, the background section plays a pivotal role in research papers.

The background of the study in research is important as it:

• Provides context: It offers readers a clear picture of the existing knowledge, helping them understand where the current research fits in.
• Highlights relevance: By detailing the reasons for the research, it underscores the study's significance and its potential impact.
• Guides the narrative: The background shapes the narrative flow of the paper, ensuring a logical progression from what's known to what the research aims to uncover.
• Enhances engagement: A well-crafted background piques the reader's interest, encouraging them to delve deeper into the research paper.
• Aids in comprehension: By setting the scenario, it aids readers in better grasping the research objectives, methodologies, and findings.

How to write the background of the study in a research paper?

The journey of presenting a compelling argument begins with the background study. This section holds the power to either captivate or lose the reader's interest.

An effectively written background not only provides context but also sets the tone for the entire research paper. It's the bridge that connects a broad topic to a specific research question, guiding readers through the logic behind the study.

But how does one craft a background of the study that resonates, informs, and engages?

Here, we’ll discuss how to write an impactful background study, ensuring your research stands out and captures the attention it deserves.

Identify the research problem

The first step is to start pinpointing the specific issue or gap you're addressing. This should be a significant and relevant problem in your field.

A well-defined problem is specific, relevant, and significant to your field. It should resonate with both experts and readers.

Here’s more on how to write an effective research problem .

Provide context

Here, you need to provide a broader perspective, illustrating how your research aligns with or contributes to the overarching context or the wider field of study. A comprehensive context is grounded in facts, offers multiple perspectives, and is relatable.

In addition to stating facts, you should weave a story that connects key concepts from the past, present, and potential future research. For instance, consider the following approach.

• Offer a brief history of the topic, highlighting major milestones or turning points that have shaped the current landscape.
• Discuss contemporary developments or current trends that provide relevant information to your research problem. This could include technological advancements, policy changes, or shifts in societal attitudes.
• Highlight the views of different stakeholders. For a topic like sustainable agriculture, this could mean discussing the perspectives of farmers, environmentalists, policymakers, and consumers.
• If relevant, compare and contrast global trends with local conditions and circumstances. This can offer readers a more holistic understanding of the topic.

Literature review

For this step, you’ll deep dive into the existing literature on the same topic. It's where you explore what scholars, researchers, and experts have already discovered or discussed about your topic.

Conducting a thorough literature review isn't just a recap of past works. To elevate its efficacy, it's essential to analyze the methods, outcomes, and intricacies of prior research work, demonstrating a thorough engagement with the existing body of knowledge.

• Instead of merely listing past research study, delve into their methodologies, findings, and limitations. Highlight groundbreaking studies and those that had contrasting results.
• Try to identify patterns. Look for recurring themes or trends in the literature. Are there common conclusions or contentious points?
• The next step would be to connect the dots. Show how different pieces of research relate to each other. This can help in understanding the evolution of thought on the topic.

By showcasing what's already known, you can better highlight the background of the study in research.

Highlight the research gap

This step involves identifying the unexplored areas or unanswered questions in the existing literature. Your research seeks to address these gaps, providing new insights or answers.

A clear research gap shows you've thoroughly engaged with existing literature and found an area that needs further exploration.

How can you efficiently highlight the research gap?

• Find the overlooked areas. Point out topics or angles that haven't been adequately addressed.
• Highlight questions that have emerged due to recent developments or changing circumstances.
• Identify areas where insights from other fields might be beneficial but haven't been explored yet.

State your objectives

Here, it’s all about laying out your game plan — What do you hope to achieve with your research? You need to mention a clear objective that’s specific, actionable, and directly tied to the research gap.

How to state your objectives?

• List the primary questions guiding your research.
• If applicable, state any hypotheses or predictions you aim to test.
• Specify what you hope to achieve, whether it's new insights, solutions, or methodologies.

Discuss the significance

This step describes your 'why'. Why is your research important? What broader implications does it have?

The significance of “why” should be both theoretical (adding to the existing literature) and practical (having real-world implications).

How do we effectively discuss the significance?

• Discuss how your research adds to the existing body of knowledge.
• Highlight how your findings could be applied in real-world scenarios, from policy changes to on-ground practices.
• Point out how your research could pave the way for further studies or open up new areas of exploration.

Summarize your points

A concise summary acts as a bridge, smoothly transitioning readers from the background to the main body of the paper. This step is a brief recap, ensuring that readers have grasped the foundational concepts.

How to summarize your study?

• Revisit the key points discussed, from the research problem to its significance.
• Prepare the reader for the subsequent sections, ensuring they understand the research's direction.

Include examples for better understanding

Research and come up with real-world or hypothetical examples to clarify complex concepts or to illustrate the practical applications of your research. Relevant examples make abstract ideas tangible, aiding comprehension.

How to include an effective example of the background of the study?

• Use past events or scenarios to explain concepts.
• Craft potential scenarios to demonstrate the implications of your findings.
• Use comparisons to simplify complex ideas, making them more relatable.

Crafting a compelling background of the study in research is about striking the right balance between providing essential context, showcasing your comprehensive understanding of the existing literature, and highlighting the unique value of your research .

While writing the background of the study, keep your readers at the forefront of your mind. Every piece of information, every example, and every objective should be geared toward helping them understand and appreciate your research.

How to avoid mistakes in the background of the study in research?

To write a well-crafted background of the study, you should be aware of the following potential research pitfalls .

• Stay away from ambiguity. Always assume that your reader might not be familiar with intricate details about your topic.
• Avoid discussing unrelated themes. Stick to what's directly relevant to your research problem.
• Ensure your background is well-organized. Information should flow logically, making it easy for readers to follow.
• While it's vital to provide context, avoid overwhelming the reader with excessive details that might not be directly relevant to your research problem.
• Ensure you've covered the most significant and relevant studies i` n your field. Overlooking key pieces of literature can make your background seem incomplete.
• Aim for a balanced presentation of facts, and avoid showing overt bias or presenting only one side of an argument.
• While academic paper often involves specialized terms, ensure they're adequately explained or use simpler alternatives when possible.
• Every claim or piece of information taken from existing literature should be appropriately cited. Failing to do so can lead to issues of plagiarism.
• Avoid making the background too lengthy. While thoroughness is appreciated, it should not come at the expense of losing the reader's interest. Maybe prefer to keep it to one-two paragraphs long.
• Especially in rapidly evolving fields, it's crucial to ensure that your literature review section is up-to-date and includes the latest research.

Example of an effective background of the study

Let's consider a topic: "The Impact of Online Learning on Student Performance." The ideal background of the study section for this topic would be as follows.

In the last decade, the rise of the internet has revolutionized many sectors, including education. Online learning platforms, once a supplementary educational tool, have now become a primary mode of instruction for many institutions worldwide. With the recent global events, such as the COVID-19 pandemic, there has been a rapid shift from traditional classroom learning to online modes, making it imperative to understand its effects on student performance.

Previous studies have explored various facets of online learning, from its accessibility to its flexibility. However, there is a growing need to assess its direct impact on student outcomes. While some educators advocate for its benefits, citing the convenience and vast resources available, others express concerns about potential drawbacks, such as reduced student engagement and the challenges of self-discipline.

This research aims to delve deeper into this debate, evaluating the true impact of online learning on student performance.

Why is this example considered as an effective background section of a research paper?

This background section example effectively sets the context by highlighting the rise of online learning and its increased relevance due to recent global events. It references prior research on the topic, indicating a foundation built on existing knowledge.

By presenting both the potential advantages and concerns of online learning, it establishes a balanced view, leading to the clear purpose of the study: to evaluate the true impact of online learning on student performance.

As we've explored, writing an effective background of the study in research requires clarity, precision, and a keen understanding of both the broader landscape and the specific details of your topic.

From identifying the research problem, providing context, reviewing existing literature to highlighting research gaps and stating objectives, each step is pivotal in shaping the narrative of your research. And while there are best practices to follow, it's equally crucial to be aware of the pitfalls to avoid.

Remember, writing or refining the background of your study is essential to engage your readers, familiarize them with the research context, and set the ground for the insights your research project will unveil.

Drawing from all the important details, insights and guidance shared, you're now in a strong position to craft a background of the study that not only informs but also engages and resonates with your readers.

Now that you've a clear understanding of what the background of the study aims to achieve, the natural progression is to delve into the next crucial component — write an effective introduction section of a research paper. Read here .

Frequently Asked Questions

The background of the study should include a clear context for the research, references to relevant previous studies, identification of knowledge gaps, justification for the current research, a concise overview of the research problem or question, and an indication of the study's significance or potential impact.

The background of the study is written to provide readers with a clear understanding of the context, significance, and rationale behind the research. It offers a snapshot of existing knowledge on the topic, highlights the relevance of the study, and sets the stage for the research questions and objectives. It ensures that readers can grasp the importance of the research and its place within the broader field of study.

The background of the study is a section in a research paper that provides context, circumstances, and history leading to the research problem or topic being explored. It presents existing knowledge on the topic and outlines the reasons that spurred the current research, helping readers understand the research's foundation and its significance in the broader academic landscape.

The number of paragraphs in the background of the study can vary based on the complexity of the topic and the depth of the context required. Typically, it might range from 3 to 5 paragraphs, but in more detailed or complex research papers, it could be longer. The key is to ensure that all relevant information is presented clearly and concisely, without unnecessary repetition.

You might also like

Smallpdf vs SciSpace: Which ChatPDF is Right for You?

ChatPDF Showdown: SciSpace Chat PDF vs. Adobe PDF Reader

Boosting Citations: A Comparative Analysis of Graphical Abstract vs. Video Abstract

Background of the Study

Ai generator.

The background of the study provides a comprehensive overview of the research problem, including the context , significance, and gaps in existing knowledge. It sets the stage for the research by outlining the historical, theoretical, and practical aspects that have led to the current investigation, highlighting the importance of addressing the identified issues.

What is the Background of a Study? 

The background of a study provides context by explaining the research problem, highlighting gaps in existing knowledge, and establishing the study’s significance. It sets the stage for the research objective , offering a foundation for understanding the study’s purpose and relevance within the broader academic discourse.

Background of the Study Format

The background of the study is a foundational section in any research paper or thesis . Here is a structured format to follow:

1. Introduction

• Briefly introduce the topic and its relevance.
• Mention the research problem or question.

2. Contextual Framework

• Provide historical background.
• Discuss relevant theories and models.
• Explain the practical context.

3. Literature Review

• Summarize key studies related to the topic.
• Highlight significant findings and their implications.
• Identify gaps in the existing literature.

4. Rationale

• Explain why the study is necessary.
• Discuss the significance and potential impact.
• Justify the research focus and scope.

5. Objectives and Research Questions

• State the primary objective of the study.
• List the specific research questions.

6. Conclusion

• Summarize the importance of the background.
• Emphasize how it sets the stage for the research.

Introduction The increasing incidence of climate change and its effects on global agriculture has raised significant concerns among researchers. This study focuses on the impact of climate change on crop yields. Contextual Framework Historically, agricultural practices have adapted to gradual climate changes. However, recent rapid shifts have outpaced these adaptations, necessitating urgent research. Theoretical models of climate adaptation provide a foundation for understanding these changes. Literature Review Recent studies show mixed results on the extent of climate change impacts on agriculture. While some regions experience reduced yields, others report minimal changes. These discrepancies highlight the need for a focused study on regional impacts. Rationale This research is crucial for developing strategies to mitigate adverse effects on agriculture. Understanding specific regional impacts can help tailor interventions, making this study highly significant for policymakers and farmers. Objectives and Research Questions To assess the impact of climate change on crop yields in the Midwest. What are the main climate factors affecting agriculture in this region? How can farmers adapt to these changes effectively? Conclusion The background of the study underscores its relevance and importance, providing a solid foundation for the research. By addressing identified gaps, this study aims to contribute valuable insights into climate change adaptation strategies in agriculture.

Background of the Study Examples

Impact of social media on academic performance, effects of urbanization on local ecosystems, role of nutrition in early childhood development.

More Background of the Study Examples

• Online Learning and Reading Skills
• Mindfulness at Work
• Parental Role in Preventing Childhood Obesity
• Green Building and Energy Efficiency
• Peer Tutoring in High Schools
• Remote Work and Work-Life Balance
• Technology in Healthcare

Background of the Study in Research Example

Background of the Study in Qualitative Research Example

Importance of Background of the Study

The background of the study is essential for several reasons:

• Context Establishment : It sets the stage for the research by outlining the historical, theoretical, and practical contexts.
• Literature Review : It provides a summary of existing literature, highlighting what is already known and identifying gaps in knowledge.
• Research Justification : It explains why the study is necessary, showcasing its relevance and significance.
• Research Direction : It guides the research questions and objectives, ensuring the study is focused and coherent.
• Foundation for Methodology : It lays the groundwork for the research methodology, explaining the choice of methods and approaches.
• Informing Stakeholders : It helps stakeholders understand the importance and potential impact of the research.

How is the Background of a Study Different From the Introduction?

The background of a study and the introduction serve distinct but complementary purposes in a research paper. Here’s how they differ:

• Provides detailed context for the research problem.
• Explains the historical, theoretical, and practical background of the topic.
• Identifies gaps in existing knowledge that the study aims to fill.
• Includes a comprehensive literature review.
• Discusses relevant theories, models, and previous research findings.
• Sets the stage for the study by explaining why it is important and necessary.
• Typically more detailed and longer than the introduction.
• Provides in-depth information to help readers understand the broader context of the research.

Introduction

• Introduces the topic and the research problem in a concise manner.
• Captures the reader’s interest and sets the stage for the rest of the paper.
• States the research objectives, questions, and sometimes hypotheses.
• Brief overview of the topic and its significance.
• Clear statement of the research problem.
• Outline of the study’s objectives and research questions.
• May include a brief mention of the methodology and scope.
• Typically shorter and more succinct than the background.
• Provides a snapshot of what the study is about without going into detailed literature review or theoretical background.

Example to Illustrate the Difference

Introduction Example : The rapid growth of social media usage among students has raised concerns about its impact on academic performance. This study aims to investigate how social media influences students’ grades and study habits. By examining different platforms and usage patterns, the research seeks to provide insights into whether social media acts as a distraction or a beneficial tool for learning. Background of the Study Example : Social media has transformed communication and information sharing, particularly among young people. Historically, educational environments have seen various technological impacts, from the introduction of computers to the widespread use of the internet. Theories of digital learning suggest both positive and negative effects of technology on education. Previous studies have shown mixed results; some indicate that social media can enhance collaborative learning and resource access, while others point to decreased academic performance due to distraction. Despite these findings, there is limited research on the long-term effects of specific social media platforms on academic outcomes. This study addresses these gaps by exploring how different types of social media usage impact student performance, aiming to provide a nuanced understanding of this contemporary issue.

Where is the Background of a Study Placed in a Research Paper? 

The background of a study is typically placed within the Introduction section of a research paper, but it can also be a separate section immediately following the introduction. Here’s a more detailed breakdown of where the background of the study can be placed:

Within the Introduction

• In many research papers, the background of the study is woven into the introduction. It provides context and justification for the research problem, leading up to the statement of the research objectives and questions.
• Starts with a general introduction to the topic.
• Provides background information and context.
• Reviews relevant literature and identifies gaps.
• States the research problem, objectives, and questions.

As a Separate Section

• In more detailed or longer research papers, the background of the study can be a standalone section that comes immediately after the introduction. This allows for a more comprehensive presentation of the context, literature review, and theoretical framework.
• Introduction : Briefly introduces the topic and states the research problem.
• Background of the Study : Provides detailed context, literature review, theoretical background, and justification for the research.
• Research Objectives and Questions : Clearly states the aims and specific questions the research seeks to answer.

How to Write a Background of the Study

Writing a background of the study involves providing a comprehensive overview of the research problem, context, and significance. Here’s a step-by-step guide on how to write an effective background of the study:

Introduce the Topic

Begin with a General Introduction : Start by introducing the broad topic to give readers an overview of the field. Example : “Social media has revolutionized communication and information sharing in the digital age.”

Provide Context

Historical Background : Explain the historical development of the topic. Example : “Historically, communication technologies have significantly influenced educational practices, from the introduction of the internet to the advent of mobile learning.” Theoretical Framework : Mention relevant theories and models. Example : “Theories such as social constructivism and digital learning provide a basis for understanding how students interact and learn through social media.”

Review Relevant Literature

Summarize Key Studies : Provide a summary of significant studies related to your topic. Example : “Previous research has shown mixed results regarding the impact of social media on academic performance. Some studies suggest that social media can be a distraction, leading to lower grades, while others indicate it can enhance learning through collaboration.” Identify Gaps in Knowledge : Highlight gaps or inconsistencies in the existing literature. Example : “Despite extensive research, there is limited understanding of the long-term effects of specific social media platforms on student performance.”

Explain the Rationale

Justify the Need for the Study : Explain why your study is necessary and important. Example : “Assessing the impact of social media on academic performance is crucial for developing effective educational strategies and policies. This study aims to fill the existing knowledge gaps by providing detailed insights into how different platforms affect student learning outcomes.”

State the Research Objectives and Questions

List the Objectives : Clearly state the main objectives of your study. Example : “The primary objectives of this study are to analyze the relationship between social media usage and academic performance and to identify the most and least beneficial platforms for students.” Pose Research Questions : Include specific research questions that guide your study. Example : “What are the main factors influencing the impact of social media on academic performance? How can students balance social media use and academic responsibilities?”

Conclude with the Importance of the Study

Summarize the Significance : Emphasize how your study will contribute to the field. Example : “This study’s findings will provide valuable insights into the role of social media in education, informing educators and policymakers on how to leverage these tools effectively to enhance student learning outcomes.”

How to avoid mistakes in writing the Background of a Study 

Avoiding mistakes in writing the background of a study involves careful planning, thorough research, and attention to detail. Here are some tips to help you avoid common mistakes:

1. Lack of Clarity and Focus

• Example : If your research is about the impact of social media on student performance, don’t delve into unrelated topics like general internet usage unless directly relevant.

2. Insufficient Literature Review

• Example : Use databases like Google Scholar, PubMed, or your institution’s library to find peer-reviewed articles and credible sources.

3. Overwhelming with Too Much Information

• Example : Summarize key studies and avoid detailed descriptions of every study you come across.

4. Failure to Identify Gaps in Knowledge

• Example : “While several studies have explored social media’s impact on general communication skills, few have examined its specific effects on academic performance among high school students.”

5. Lack of Theoretical Framework

• Example : “The study is grounded in social constructivism, which suggests that learning occurs through social interactions, making it relevant to examine how social media platforms facilitate these interactions.”

6. Inadequate Justification for the Study

• Example : “Understanding the impact of social media on academic performance is crucial for developing effective educational strategies and policies.”

7. Poor Organization and Structure

• Example : Use clear headings like “Introduction,” “Contextual Framework,” “Literature Review,” “Rationale,” and “Research Objectives and Questions.”

8. Using Jargon and Complex Language

• Example : Instead of “The pedagogical implications of digital media necessitate a paradigmatic shift,” say “Digital media impacts teaching methods, requiring changes in how we educate.”

9. Ignoring the Research Objectives and Questions

• Example : “This background review highlights the need to investigate how different social media platforms affect high school students’ study habits, directly addressing the research questions outlined.”

10. Neglecting to Update References

• Example : Instead of relying solely on sources from over a decade ago, incorporate recent studies that reflect current trends and findings.

What is the background of the study?

The background of the study provides context, explains the research problem, reviews relevant literature, and identifies gaps the study aims to fill.

Why is the background of the study important?

It establishes the context and significance of the research, justifies the study, and helps readers understand the broader academic landscape and gaps the research addresses.

How does the background of the study differ from the introduction?

The background provides detailed context and literature review, while the introduction briefly presents the research problem, objectives, and significance.

What should be included in the background of the study?

Include historical context, theoretical framework, literature review, gaps in knowledge, and the rationale for the study.

Where is the background of the study placed in a research paper?

It is typically integrated within the introduction or presented as a separate section following the introduction.

How long should the background of the study be?

The length varies, but it should be detailed enough to provide context and justification, typically a few paragraphs to several pages.

How do you write a strong background of the study?

Conduct thorough research, organize logically, include relevant theories and studies, identify gaps, and justify the research’s importance.

Can the background of the study include preliminary data?

Yes, including preliminary data can strengthen the background by demonstrating initial findings and supporting the research rationale.

How do you identify gaps in the literature?

Conduct a comprehensive literature review, compare findings, and note inconsistencies, unexplored areas, or outdated research that your study will address.

Should the background of the study be written in chronological order?

Not necessarily. Organize logically by themes, concepts, or research gaps rather than strictly chronologically to provide a coherent context for your study.

Text prompt

• Instructive
• Professional

10 Examples of Public speaking

20 Examples of Gas lighting

Academia.edu no longer supports Internet Explorer.

To browse Academia.edu and the wider internet faster and more securely, please take a few seconds to  upgrade your browser .

Enter the email address you signed up with and we'll email you a reset link.

• We're Hiring!
• Help Center

PROCEDURE FOR WRITING A BACKGROUND STUDY FOR A RESEARCH PAPER - WITH A PRACTICAL EXAMPLE BY DR BENARD LANGO

2020, DR BENARD LANGO RESEARCH SCHOOL

Many research documents when reviewed wholesomely in most instances fails the background study test as the authors either presume it is a section notes on the research or just a section to ensure is fully written with materials related to the research area. It is important to note that research background study will define the relevance of the study topic and whether its intention to contribute to the area of knowledge is relevant. In order to be able to write background study the study area commonly described by the study topic will be of at most importance. From the topic one should be able to derive the both the independent and dependent variables to be able to follow this structure of writing a background for the study.

Related Papers

Journal of Universal College of Medical Sciences

bishal joshi

INTRODUCTION A research paper is a part of academic writing where there is a gathering of information from different sources. It is multistep process. Selection of title is the most important part of research writing. The title which is interesting should be chosen for the research purpose. All the related information is gathered and the title for research is synthesized. After thorough understanding and developing the title, the preliminary outline is made which maintains the logical path for its exploration. After preliminary research, proper research work is started with collection of previous resources which is then organized and important points are noted. Then research paper is written by referring to outlines, notes, articles, journals and books. The research paper should be well structured containing core parts like introduction, material and methods, results and disscussion and important additional parts like title, abstract, references.

Sher Singh Bhakar

Organisation of Book: The book is organized into two parts. Part one starts with thinking critically about research, explains what is (and isn’t) research, explains how to properly use research in your writing to make your points, introduces a series of writing exercises designed to help students to think about and write effective research papers. Instead of explaining how to write a single “research paper,” The Process of Research Writing part of the book breaks down the research process into many smaller and easier-to manage parts like what is a research paper, starting steps for writing research papers, writing conceptual understanding and review of literature, referencing including various styles of referencing, writing research methodology and results including interpretations, writing implications and limitations of research and what goes into conclusions. Part two contains sample research articles to demonstrate the application of techniques and methods of writing good resear...

Kevin O'Donnell

Maluk Them John

Jeff Immanuel

Ana Stefanelli

Scientific Research Journal of Clinical and Medical Sciences

Ahmed Alkhaqani

Background: Confusion about elements of a research paper is common among students. The key to writing a good research paper is to know these common elements and their definitions. Maybe find that writing a research paper is not as easy as it seems. There are many parts and steps to the process, and it can be hard to figure out what needs to do and when. Objective: This article aims to teach these common aspects of a research paper to avoid common mistakes while drafting own. Conclusion: Each section of the research paper serves a distinct purpose and highlights a different aspect of the research. However, before starting drafting the manuscript, having a clear understanding of each section's purposes will help avoid mistakes.

IAA Journal of Applied Sciences

EZE V A L H Y G I N U S UDOKA

Many young researchers find it difficult to write a good and quality research thesis/article because they are not prone to article writing ethics and training. Yet, a thesis/publication is often vital and paramount for career advancement, grants, academic qualifications and others. This research work described the basics and systematic steps to follow in writing a good scientific thesis/article. This research also outlined the main sections that an average thesis/article should contain, the elements that should appear in each section, the systematic approaches in writing research, the characteristics of a good thesis/article, the attributes of a good research thesis/article, qualities of a good researcher and finally the ethics guiding research.

Loading Preview

Sorry, preview is currently unavailable. You can download the paper by clicking the button above.

RELATED PAPERS

Langley, BC: Trinity Western University. …

Paul T P Wong

Stevejobs.education

Dr. David Annan

Muhammad Obaidullah

Hossein Saiedian

Putri Ramadhani

abasynuniv.edu.pk

Flora Maleki

Renaldi Bimantoro

British journal of community nursing

Keith Meadows

Holuphumiee Adegbaju

Maamar Missoum

christopher Boateng

Faith Gatune

nikhil arora

Alex Galarosa

Akhil Rashinkar

Lawrence Rudner

IJAR Indexing

UncleChew Bah

almina ocampo

RELATED TOPICS

•   We're Hiring!
•   Help Center
• Find new research papers in:
• Health Sciences
• Earth Sciences
• Cognitive Science
• Mathematics
• Computer Science
• Academia ©2024

PHILO-notes

Free Online Learning Materials

How to Write the Background of the Study in Research (Part 1)

Background of the Study in Research: Definition and the Core Elements it Contains

Before we embark on a detailed discussion on how to write the background of the study of your proposed research or thesis, it is important to first discuss its meaning and the core elements that it should contain. This is obviously because understanding the nature of the background of the study in research and knowing exactly what to include in it allow us to have both greater control and clear direction of the writing process.

So, what really is the background of the study and what are the core elements that it should contain?

The background of the study, which usually forms the first section of the introduction to a research paper or thesis, provides the overview of the study. In other words, it is that section of the research paper or thesis that establishes the context of the study. Its main function is to explain why the proposed research is important and essential to understanding the main aspects of the study.

The background of the study, therefore, is the section of the research paper or thesis that identifies the problem or gap of the study that needs to addressed and justifies the need for conducting the study. It also articulates the main goal of the study and the thesis statement, that is, the main claim or argument of the paper.

Given this brief understanding of the background of the study, we can anticipate what readers or thesis committee members expect from it. As we can see, the background of the study should contain the following major points:

1) brief discussion on what is known about the topic under investigation; 2) An articulation of the research gap or problem that needs to be addressed; 3) What the researcher would like to do or aim to achieve in the study ( research goal); 4) The thesis statement, that is, the main argument or contention of the paper (which also serves as the reason why the researcher would want to pursue the study); 5) The major significance or contribution of the study to a particular discipline; and 6) Depending on the nature of the study, an articulation of the hypothesis of the study.

Thus, when writing the background of the study, you should plan and structure it based on the major points just mentioned. With this, you will have a clear picture of the flow of the tasks that need to be completed in writing this section of your research or thesis proposal.

Now, how do you go about writing the background of the study in your proposed research or thesis?

The next lessons will address this question.

How to Write the Opening Paragraphs of the Background of the Study?

To begin with, let us assume that you already have conducted a preliminary research on your chosen topic, that is, you already have read a lot of literature and gathered relevant information for writing the background of your study. Let us also assume that you already have identified the gap of your proposed research and have already developed the research questions and thesis statement. If you have not yet identified the gap in your proposed research, you might as well go back to our lesson on how to identify a research gap.

So, we will just put together everything that you have researched into a background of the study (assuming, again, that you already have the necessary information). But in this lesson, let’s just focus on writing the opening paragraphs.

It is important to note at this point that there are different styles of writing the background of the study. Hence, what I will be sharing with you here is not just “the” only way of writing the background of the study. As a matter of fact, there is no “one-size-fits-all” style of writing this part of the research or thesis. At the end of the day, you are free to develop your own. However, whatever style it would be, it always starts with a plan which structures the writing process into stages or steps. The steps that I will share with below are just some of the most effective ways of writing the background of the study in research.

So, let’s begin.

It is always a good idea to begin the background of your study by giving an overview of your research topic. This may include providing a definition of the key concepts of your research or highlighting the main developments of the research topic.

Let us suppose that the topic of your study is the “lived experiences of students with mathematical anxiety”.

Here, you may start the background of your study with a discussion on the meaning, nature, and dynamics of the term “mathematical anxiety”. The reason for this is too obvious: “mathematical anxiety” is a highly technical term that is specific to mathematics. Hence, this term is not readily understandable to non-specialists in this field.

So, you may write the opening paragraph of your background of the study with this:

“Mathematical anxiety refers to the individual’s unpleasant emotional mood responses when confronted with a mathematical situation.”

Since you do not invent the definition of the term “mathematical anxiety”, then you need to provide a citation to the source of the material from which you are quoting. For example, you may now say:

“Mathematical anxiety refers to the individual’s unpleasant emotional mood responses when confronted with a mathematical situation (Eliot, 2020).”

And then you may proceed with the discussion on the nature and dynamics of the term “mathematical anxiety”. You may say:

“Lou (2019) specifically identifies some of the manifestations of this type of anxiety, which include, but not limited to, depression, helplessness, nervousness and fearfulness in doing mathematical and numerical tasks.”

After explaining to your readers the meaning, nature, and dynamics (as well as some historical development if you wish to) of the term “mathematical anxiety”, you may now proceed to showing the problem or gap of the study. As you may already know, the research gap is the problem that needs to be addressed in the study. This is important because no research activity is possible without the research gap.

Let us suppose that your research problem or gap is: “Mathematical anxiety can negatively affect not just the academic achievement of the students but also their future career plans and total well-being. Also, there are no known studies that deal with the mathematical anxiety of junior high school students in New Zealand.” With this, you may say:

“If left unchecked, as Shapiro (2019) claims, this problem will expand and create a total avoidance pattern on the part of the students, which can be expressed most visibly in the form of cutting classes and habitual absenteeism. As we can see, this will negatively affect the performance of students in mathematics. In fact, the study conducted by Luttenberger and Wimmer (2018) revealed that the outcomes of mathematical anxiety do not only negatively affect the students’ performance in math-related situations but also their future career as professionals. Without a doubt, therefore, mathematical anxiety is a recurring problem for many individuals which will negatively affect the academic success and future career of the student.”

Now that you already have both explained the meaning, nature, and dynamics of the term “mathematical anxiety” and articulated the gap of your proposed research, you may now state the main goal of your study. You may say:

“Hence, it is precisely in this context that the researcher aims to determine the lived experiences of those students with mathematical anxiety. In particular, this proposed thesis aims to determine the lived experiences of the junior high school students in New Zealand and identify the factors that caused them to become disinterested in mathematics.”

Please note that you should not end the first paragraph of your background of the study with the articulation of the research goal. You also need to articulate the “thesis statement”, which usually comes after the research goal. As is well known, the thesis statement is the statement of your argument or contention in the study. It is more of a personal argument or claim of the researcher, which specifically highlights the possible contribution of the study. For example, you may say:

“The researcher argues that there is a need to determine the lived experiences of these students with mathematical anxiety because knowing and understanding the difficulties and challenges that they have encountered will put the researcher in the best position to offer some alternatives to the problem. Indeed, it is only when we have performed some kind of a ‘diagnosis’ that we can offer practicable solutions to the problem. And in the case of the junior high school students in New Zealand who are having mathematical anxiety, determining their lived experiences as well as identifying the factors that caused them to become disinterested in mathematics are the very first steps in addressing the problem.”

If we combine the bits and pieces that we have written above, we can now come up with the opening paragraphs of your background of the study, which reads:

As we can see, we can find in the first paragraph 5 essential elements that must be articulated in the background of the study, namely:

1) A brief discussion on what is known about the topic under investigation; 2) An articulation of the research gap or problem that needs to be addressed; 3) What the researcher would like to do or aim to achieve in the study (research goal); 4) The thesis statement , that is, the main argument or claim of the paper; and 5) The major significance or contribution of the study to a particular discipline. So, that’s how you write the opening paragraphs of your background of the study. The next lesson will talk about writing the body of the background of the study.

How to Write the Body of the Background of the Study?

If we liken the background of the study to a sitting cat, then the opening paragraphs that we have completed in the previous lesson would just represent the head of the cat.

This means we still have to write the body (body of the cat) and the conclusion (tail). But how do we write the body of the background of the study? What should be its content?

Truly, this is one of the most difficult challenges that fledgling scholars faced. Because they are inexperienced researchers and didn’t know what to do next, they just wrote whatever they wished to write. Fortunately, this is relatively easy if they know the technique.

One of the best ways to write the body of the background of the study is to attack it from the vantage point of the research gap. If you recall, when we articulated the research gap in the opening paragraphs, we made a bold claim there, that is, there are junior high school students in New Zealand who are experiencing mathematical anxiety. Now, you have to remember that a “statement” remains an assumption until you can provide concrete proofs to it. This is what we call the “epistemological” aspect of research. As we may already know, epistemology is a specific branch of philosophy that deals with the validity of knowledge. And to validate knowledge is to provide concrete proofs to our statements. Hence, the reason why we need to provide proofs to our claim that there are indeed junior high school students in New Zealand who are experiencing mathematical anxiety is the obvious fact that if there are none, then we cannot proceed with our study. We have no one to interview with in the first. In short, we don’t have respondents.

The body of the background of the study, therefore, should be a presentation and articulation of the proofs to our claim that indeed there are junior high school students in New Zealand who are experiencing mathematical anxiety. Please note, however, that this idea is true only if you follow the style of writing the background of the study that I introduced in this course.

So, how do we do this?

One of the best ways to do this is to look for literature on mathematical anxiety among junior high school students in New Zealand and cite them here. However, if there are not enough literature on this topic in New Zealand, then we need to conduct initial interviews with these students or make actual classroom observations and record instances of mathematical anxiety among these students. But it is always a good idea if we combine literature review with interviews and actual observations.

Assuming you already have the data, then you may now proceed with the writing of the body of your background of the study. For example, you may say:

“According to records and based on the researcher’s firsthand experience with students in some junior high schools in New Zealand, indeed, there are students who lost interest in mathematics. For one, while checking the daily attendance and monitoring of the students, it was observed that some of them are not always attending classes in mathematics but are regularly attending the rest of the required subjects.”

After this sentence, you may insert some literature that will support this position. For example, you may say:

“As a matter of fact, this phenomenon is also observed in the work of Estonanto. In his study titled ‘Impact of Math Anxiety on Academic Performance in Pre-Calculus of Senior High School’, Estonanto (2019) found out that, inter alia, students with mathematical anxiety have the tendency to intentionally prioritize other subjects and commit habitual tardiness and absences.”

Then you may proceed saying:

“With this initial knowledge in mind, the researcher conducted initial interviews with some of these students. The researcher learned that one student did not regularly attend his math subject because he believed that he is not good in math and no matter how he listens to the topic he will not learn.”

Then you may say:

“Another student also mentioned that she was influenced by her friends’ perception that mathematics is hard; hence, she avoids the subject. Indeed, these are concrete proofs that there are some junior high school students in New Zealand who have mathematical anxiety. As already hinted, “disinterest” or the loss of interest in mathematics is one of the manifestations of a mathematical anxiety.”

If we combine what we have just written above, then we can have the first two paragraphs of the body of our background of the study. It reads:

“According to records and based on the researcher’s firsthand experience with students in some junior high schools in New Zealand, indeed there are students who lost interest in mathematics. For one, while checking the daily attendance and monitoring of the students, it was observed that some of them are not always attending classes in mathematics but are regularly attending the rest of the required subjects. As a matter of fact, this phenomenon is also observed in the work of Estonanto. In his study titled ‘Impact of Math Anxiety on Academic Performance in Pre-Calculus of Senior High School’, Estonanto (2019) found out that, inter alia, students with mathematical anxiety have the tendency to intentionally prioritize other subjects and commit habitual tardiness and absences.

With this initial knowledge in mind, the researcher conducted initial interviews with some of these students. The researcher learned that one student did not regularly attend his math subject because he believed that he is not good in math and no matter how he listens to the topic he will not learn. Another student also mentioned that she was influenced by her friends’ perception that mathematics is hard; hence, she avoids the subject. Indeed, these are concrete proofs that there are some junior high school students in New Zealand who have mathematical anxiety. As already hinted, “disinterest” or the loss of interest in mathematics is one of the manifestations of a mathematical anxiety.”

And then you need validate this observation by conducting another round of interview and observation in other schools. So, you may continue writing the body of the background of the study with this:

“To validate the information gathered from the initial interviews and observations, the researcher conducted another round of interview and observation with other junior high school students in New Zealand.”

“On the one hand, the researcher found out that during mathematics time some students felt uneasy; in fact, they showed a feeling of being tensed or anxious while working with numbers and mathematical problems. Some were even afraid to seat in front, while some students at the back were secretly playing with their mobile phones. These students also show remarkable apprehension during board works like trembling hands, nervous laughter, and the like.”

Then provide some literature that will support your position. You may say:

“As Finlayson (2017) corroborates, emotional symptoms of mathematical anxiety involve feeling of helplessness, lack of confidence, and being nervous for being put on the spot. It must be noted that these occasionally extreme emotional reactions are not triggered by provocative procedures. As a matter of fact, there are no personally sensitive questions or intentional manipulations of stress. The teacher simply asked a very simple question, like identifying the parts of a circle. Certainly, this observation also conforms with the study of Ashcraft (2016) when he mentions that students with mathematical anxiety show a negative attitude towards math and hold self-perceptions about their mathematical abilities.”

And then you proceed:

“On the other hand, when the class had their other subjects, the students show a feeling of excitement. They even hurried to seat in front and attentively participating in the class discussion without hesitation and without the feeling of being tensed or anxious. For sure, this is another concrete proof that there are junior high school students in New Zealand who have mathematical anxiety.”

To further prove the point that there indeed junior high school students in New Zealand who have mathematical anxiety, you may solicit observations from other math teachers. For instance, you may say:

“The researcher further verified if the problem is also happening in other sections and whether other mathematics teachers experienced the same observation that the researcher had. This validation or verification is important in establishing credibility of the claim (Buchbinder, 2016) and ensuring reliability and validity of the assertion (Morse et al., 2002). In this regard, the researcher attempted to open up the issue of math anxiety during the Departmentalized Learning Action Cell (LAC), a group discussion of educators per quarter, with the objective of ‘Teaching Strategies to Develop Critical Thinking of the Students’. During the session, one teacher corroborates the researcher’s observation that there are indeed junior high school students in New Zealand who have mathematical anxiety. The teacher pointed out that truly there were students who showed no extra effort in mathematics class in addition to the fact that some students really avoided the subject. In addition, another math teacher expressed her frustrations about these students who have mathematical anxiety. She quipped: “How can a teacher develop the critical thinking skills or ability of the students if in the first place these students show avoidance and disinterest in the subject?’.”

Again, if we combine what we have just written above, then we can now have the remaining parts of the body of the background of the study. It reads:

So, that’s how we write the body of the background of the study in research . Of course, you may add any relevant points which you think might amplify your content. What is important at this point is that you now have a clear idea of how to write the body of the background of the study.

How to Write the Concluding Part of the Background of the Study?

Since we have already completed the body of our background of the study in the previous lesson, we may now write the concluding paragraph (the tail of the cat). This is important because one of the rules of thumb in writing is that we always put a close to what we have started.

It is important to note that the conclusion of the background of the study is just a rehashing of the research gap and main goal of the study stated in the introductory paragraph, but framed differently. The purpose of this is just to emphasize, after presenting the justifications, what the study aims to attain and why it wants to do it. The conclusion, therefore, will look just like this:

“Given the above discussion, it is evident that there are indeed junior high school students in New Zealand who are experiencing mathematical anxiety. And as we can see, mathematical anxiety can negatively affect not just the academic achievement of the students but also their future career plans and total well-being. Again, it is for this reason that the researcher attempts to determine the lived experiences of those junior high school students in New Zealand who are experiencing a mathematical anxiety.”

If we combine all that we have written from the very beginning, the entire background of the study would now read:

If we analyze the background of the study that we have just completed, we can observe that in addition to the important elements that it should contain, it has also addressed other important elements that readers or thesis committee members expect from it.

On the one hand, it provides the researcher with a clear direction in the conduct of the study. As we can see, the background of the study that we have just completed enables us to move in the right direction with a strong focus as it has set clear goals and the reasons why we want to do it. Indeed, we now exactly know what to do next and how to write the rest of the research paper or thesis.

On the other hand, most researchers start their research with scattered ideas and usually get stuck with how to proceed further. But with a well-written background of the study, just as the one above, we have decluttered and organized our thoughts. We have also become aware of what have and have not been done in our area of study, as well as what we can significantly contribute in the already existing body of knowledge in this area of study.

Please note, however, as I already mentioned previously, that the model that I have just presented is only one of the many models available in textbooks and other sources. You are, of course, free to choose your own style of writing the background of the study. You may also consult your thesis supervisor for some guidance on how to attack the writing of your background of the study.

Lastly, and as you may already know, universities around the world have their own thesis formats. Hence, you should follow your university’s rules on the format and style in writing your research or thesis. What is important is that with the lessons that you learned in this course, you can now easily write the introductory part of your thesis, such as the background of the study.

Global site navigation

• Celebrity biographies
• Messages - Wishes - Quotes
• TV and Movies
• Fashion and style
• Music and singers
• Capital Market
• Celebrities
• Relationships

Local editions

• Habari za Kenya Swahili

How to write background of the study in research proposal

Researchers understand that the background of the study of any research paper is one of the most crucial components. It summarizes all details in the research paper, giving a first-time reader a chance to figure out what the research is about. It follows a strict format and includes sources used, previous research reviews, as well as a conclusion that offers potential solutions to the problem being researched.

Knowing how to write background of the study properly is a skill every writer needs to master. First things first though, what is the meaning of background of the study in research? It is the information that a reader needs to increase his awareness of the topic an essay is going to explain. It is the part that comes immediately after the hook or attention grabber.

The background of the study meaning is simple and straightforward, as it serves as a comprehensive introduction of the entire research paper.

How to write a report (with samples)

It shows that the chosen research questions or thesis statements are significant and solvable. The introduction and background of the study must be written properly and with clarity.

Avoid choosing hypothetical questions. A good research paper aims at solving a problem that exists in a community, with the intention of providing a solution that works.

Understanding the background of the study in research proposal

Do you know how to write background of the study in research proposal? The process is easy. Follow the simple steps below and stick to the format described.

What does a research background example include?

A good research background has the following components:

• It should have reviews of the area being researched
• It should have currently available information about the problem of the study
• It should capture the previous studies on the issue
• It must indicate the history of the issue of the study from previous researches done on the subject

How to cite a website in MLA or APA

READ ALSO: How to write an informal letter: Format and samples

Steps to follow when writing the background of the study

It is important to figure out what goes into the background of the study and the procedures involved when writing.

Typically, the background of a study includes a number of things starting from the review of the existing literature on the area of your research, which will, in turn, lead up to your topic.

After discussing the contribution of other past researchers in the field, you can then identify gaps in understanding the subject. This means focusing on areas that have not been addressed in these previous studies.

Lastly, you should explain how your study will address these gaps and the contributions it makes to the existing knowledge in the field.

This, therefore, means that a good background of the study example follows the steps below.

Minute writing: tips, examples and templates

1. Formulating the thesis

Any good example of background of the study in research paper captures the thesis statement clearly. Note that most of the issues under investigation will be unclear when you start working on your paper. Focus on putting together ideas and preliminary research.

Come up with a thesis on a specific topic whose relevance to the study is clear.

2. Read and gather relevant information

Coming up with a thesis statement involves intensive research and reading. Write down all you discover and make notes. Stick to sources that are considered reliable, making sure you reference and keep track of everything.

3. Developing the research question or thesis statement

You have to give an authoritative opinion or position about the research you are conducting from your reading and research.

4. Complete your research

At this point, use the questions and thesis statement as a guide to complete your research. Focus on using sources that are relevant to the specific requirements of your questions.

Different types of feasibility study

5. Create a good structure

This is the point where you start creating several varied sections. Discuss key issues, major findings, as well as controversies that are linked to the research issue. You should also have a section with the evaluation and conclusion.

6. Identify areas requiring further studies

Come up with possible solutions, especially those that have not been factored in or applied before. In most cases, the areas will be part of the recommendations you make as you conclude.

7. Proofread your work

A good research background example should be thoroughly proofread. Read through the paper severally and also have someone else go through the completed write-up, just in case you missed any mistakes.

Writing form

Now that you know what steps to follow when writing the background, the next thing is to understand the format to use.

• Start by giving a general overview of the thesis statement. Start by introducing all the ideas you intend to discuss.
• The next step would be to give detailed and specific methodologies used in the research. The length of this section will be dependent on the research topic and questions. However, it is expected to be several pages.
• Remember to cite sources as needed. This will help avoid any cases of plagiarism.
• You can then introduce the experiment you intend to carry out by describing the methodology to be used. The reasons for choosing the method should be clearly described. Explain why the chosen method is best preferred as opposed to others, describing the objectives of the chosen methods.

How to write a work plan (with template and samples)

A good background of the study sample has all the above-mentioned components. Including the details in the background of the study ensures your research is complete.

READ ALSO :

• Simple character bio template
• Simple call sheet template
• How to write an informal letter: Format and samples

Source: TUKO.co.ke

Jackline Wangare (Lifestyle writer) Jackline Simwa is a content writer at Tuko.co.ke, where she has worked since mid-2021. She tackles diverse topics, including finance, entertainment, sports, and lifestyle. Previously, she worked at The Campanile at Kenyatta University. She has more than five years in writing. Jackline graduated with a Bachelor’s degree in Economics (2019) and a Diploma in Marketing (2015) from Kenyatta University. In 2023, Simwa finished the AFP course on Digital Investigation Techniques and Google News Initiative course in 2024. Email: [email protected].

• Privacy Policy

Home » Significance of the Study – Examples and Writing Guide

Significance of the Study – Examples and Writing Guide

Table of Contents

Significance of the Study

Definition:

Significance of the study in research refers to the potential importance, relevance, or impact of the research findings. It outlines how the research contributes to the existing body of knowledge, what gaps it fills, or what new understanding it brings to a particular field of study.

In general, the significance of a study can be assessed based on several factors, including:

• Originality : The extent to which the study advances existing knowledge or introduces new ideas and perspectives.
• Practical relevance: The potential implications of the study for real-world situations, such as improving policy or practice.
• Theoretical contribution: The extent to which the study provides new insights or perspectives on theoretical concepts or frameworks.
• Methodological rigor : The extent to which the study employs appropriate and robust methods and techniques to generate reliable and valid data.
• Social or cultural impact : The potential impact of the study on society, culture, or public perception of a particular issue.

Types of Significance of the Study

The significance of the Study can be divided into the following types:

Theoretical Significance

Theoretical significance refers to the contribution that a study makes to the existing body of theories in a specific field. This could be by confirming, refuting, or adding nuance to a currently accepted theory, or by proposing an entirely new theory.

Practical Significance

Practical significance refers to the direct applicability and usefulness of the research findings in real-world contexts. Studies with practical significance often address real-life problems and offer potential solutions or strategies. For example, a study in the field of public health might identify a new intervention that significantly reduces the spread of a certain disease.

Significance for Future Research

This pertains to the potential of a study to inspire further research. A study might open up new areas of investigation, provide new research methodologies, or propose new hypotheses that need to be tested.

How to Write Significance of the Study

Here’s a guide to writing an effective “Significance of the Study” section in research paper, thesis, or dissertation:

• Background : Begin by giving some context about your study. This could include a brief introduction to your subject area, the current state of research in the field, and the specific problem or question your study addresses.
• Identify the Gap : Demonstrate that there’s a gap in the existing literature or knowledge that needs to be filled, which is where your study comes in. The gap could be a lack of research on a particular topic, differing results in existing studies, or a new problem that has arisen and hasn’t yet been studied.
• State the Purpose of Your Study : Clearly state the main objective of your research. You may want to state the purpose as a solution to the problem or gap you’ve previously identified.
• Contributes to the existing body of knowledge.
• Addresses a significant research gap.
• Offers a new or better solution to a problem.
• Impacts policy or practice.
• Leads to improvements in a particular field or sector.
• Identify Beneficiaries : Identify who will benefit from your study. This could include other researchers, practitioners in your field, policy-makers, communities, businesses, or others. Explain how your findings could be used and by whom.
• Future Implications : Discuss the implications of your study for future research. This could involve questions that are left open, new questions that have been raised, or potential future methodologies suggested by your study.

Significance of the Study in Research Paper

The Significance of the Study in a research paper refers to the importance or relevance of the research topic being investigated. It answers the question “Why is this research important?” and highlights the potential contributions and impacts of the study.

The significance of the study can be presented in the introduction or background section of a research paper. It typically includes the following components:

• Importance of the research problem: This describes why the research problem is worth investigating and how it relates to existing knowledge and theories.
• Potential benefits and implications: This explains the potential contributions and impacts of the research on theory, practice, policy, or society.
• Originality and novelty: This highlights how the research adds new insights, approaches, or methods to the existing body of knowledge.
• Scope and limitations: This outlines the boundaries and constraints of the research and clarifies what the study will and will not address.

Suppose a researcher is conducting a study on the “Effects of social media use on the mental health of adolescents”.

The significance of the study may be:

“The present study is significant because it addresses a pressing public health issue of the negative impact of social media use on adolescent mental health. Given the widespread use of social media among this age group, understanding the effects of social media on mental health is critical for developing effective prevention and intervention strategies. This study will contribute to the existing literature by examining the moderating factors that may affect the relationship between social media use and mental health outcomes. It will also shed light on the potential benefits and risks of social media use for adolescents and inform the development of evidence-based guidelines for promoting healthy social media use among this population. The limitations of this study include the use of self-reported measures and the cross-sectional design, which precludes causal inference.”

Significance of the Study In Thesis

The significance of the study in a thesis refers to the importance or relevance of the research topic and the potential impact of the study on the field of study or society as a whole. It explains why the research is worth doing and what contribution it will make to existing knowledge.

For example, the significance of a thesis on “Artificial Intelligence in Healthcare” could be:

• With the increasing availability of healthcare data and the development of advanced machine learning algorithms, AI has the potential to revolutionize the healthcare industry by improving diagnosis, treatment, and patient outcomes. Therefore, this thesis can contribute to the understanding of how AI can be applied in healthcare and how it can benefit patients and healthcare providers.
• AI in healthcare also raises ethical and social issues, such as privacy concerns, bias in algorithms, and the impact on healthcare jobs. By exploring these issues in the thesis, it can provide insights into the potential risks and benefits of AI in healthcare and inform policy decisions.
• Finally, the thesis can also advance the field of computer science by developing new AI algorithms or techniques that can be applied to healthcare data, which can have broader applications in other industries or fields of research.

Significance of the Study in Research Proposal

The significance of a study in a research proposal refers to the importance or relevance of the research question, problem, or objective that the study aims to address. It explains why the research is valuable, relevant, and important to the academic or scientific community, policymakers, or society at large. A strong statement of significance can help to persuade the reviewers or funders of the research proposal that the study is worth funding and conducting.

Here is an example of a significance statement in a research proposal:

Title : The Effects of Gamification on Learning Programming: A Comparative Study

Significance Statement:

This proposed study aims to investigate the effects of gamification on learning programming. With the increasing demand for computer science professionals, programming has become a fundamental skill in the computer field. However, learning programming can be challenging, and students may struggle with motivation and engagement. Gamification has emerged as a promising approach to improve students’ engagement and motivation in learning, but its effects on programming education are not yet fully understood. This study is significant because it can provide valuable insights into the potential benefits of gamification in programming education and inform the development of effective teaching strategies to enhance students’ learning outcomes and interest in programming.

Examples of Significance of the Study

Here are some examples of the significance of a study that indicates how you can write this into your research paper according to your research topic:

Research on an Improved Water Filtration System : This study has the potential to impact millions of people living in water-scarce regions or those with limited access to clean water. A more efficient and affordable water filtration system can reduce water-borne diseases and improve the overall health of communities, enabling them to lead healthier, more productive lives.

Study on the Impact of Remote Work on Employee Productivity : Given the shift towards remote work due to recent events such as the COVID-19 pandemic, this study is of considerable significance. Findings could help organizations better structure their remote work policies and offer insights on how to maximize employee productivity, wellbeing, and job satisfaction.

Investigation into the Use of Solar Power in Developing Countries : With the world increasingly moving towards renewable energy, this study could provide important data on the feasibility and benefits of implementing solar power solutions in developing countries. This could potentially stimulate economic growth, reduce reliance on non-renewable resources, and contribute to global efforts to combat climate change.

Research on New Learning Strategies in Special Education : This study has the potential to greatly impact the field of special education. By understanding the effectiveness of new learning strategies, educators can improve their curriculum to provide better support for students with learning disabilities, fostering their academic growth and social development.

Examination of Mental Health Support in the Workplace : This study could highlight the impact of mental health initiatives on employee wellbeing and productivity. It could influence organizational policies across industries, promoting the implementation of mental health programs in the workplace, ultimately leading to healthier work environments.

Evaluation of a New Cancer Treatment Method : The significance of this study could be lifesaving. The research could lead to the development of more effective cancer treatments, increasing the survival rate and quality of life for patients worldwide.

When to Write Significance of the Study

The Significance of the Study section is an integral part of a research proposal or a thesis. This section is typically written after the introduction and the literature review. In the research process, the structure typically follows this order:

• Title – The name of your research.
• Abstract – A brief summary of the entire research.
• Introduction – A presentation of the problem your research aims to solve.
• Literature Review – A review of existing research on the topic to establish what is already known and where gaps exist.
• Significance of the Study – An explanation of why the research matters and its potential impact.

In the Significance of the Study section, you will discuss why your study is important, who it benefits, and how it adds to existing knowledge or practice in your field. This section is your opportunity to convince readers, and potentially funders or supervisors, that your research is valuable and worth undertaking.

Advantages of Significance of the Study

The Significance of the Study section in a research paper has multiple advantages:

• Establishes Relevance: This section helps to articulate the importance of your research to your field of study, as well as the wider society, by explicitly stating its relevance. This makes it easier for other researchers, funders, and policymakers to understand why your work is necessary and worth supporting.
• Guides the Research: Writing the significance can help you refine your research questions and objectives. This happens as you critically think about why your research is important and how it contributes to your field.
• Attracts Funding: If you are seeking funding or support for your research, having a well-written significance of the study section can be key. It helps to convince potential funders of the value of your work.
• Opens up Further Research: By stating the significance of the study, you’re also indicating what further research could be carried out in the future, based on your work. This helps to pave the way for future studies and demonstrates that your research is a valuable addition to the field.
• Provides Practical Applications: The significance of the study section often outlines how the research can be applied in real-world situations. This can be particularly important in applied sciences, where the practical implications of research are crucial.
• Enhances Understanding: This section can help readers understand how your study fits into the broader context of your field, adding value to the existing literature and contributing new knowledge or insights.

Limitations of Significance of the Study

The Significance of the Study section plays an essential role in any research. However, it is not without potential limitations. Here are some that you should be aware of:

• Subjectivity: The importance and implications of a study can be subjective and may vary from person to person. What one researcher considers significant might be seen as less critical by others. The assessment of significance often depends on personal judgement, biases, and perspectives.
• Predictability of Impact: While you can outline the potential implications of your research in the Significance of the Study section, the actual impact can be unpredictable. Research doesn’t always yield the expected results or have the predicted impact on the field or society.
• Difficulty in Measuring: The significance of a study is often qualitative and can be challenging to measure or quantify. You can explain how you think your research will contribute to your field or society, but measuring these outcomes can be complex.
• Possibility of Overstatement: Researchers may feel pressured to amplify the potential significance of their study to attract funding or interest. This can lead to overstating the potential benefits or implications, which can harm the credibility of the study if these results are not achieved.
• Overshadowing of Limitations: Sometimes, the significance of the study may overshadow the limitations of the research. It is important to balance the potential significance with a thorough discussion of the study’s limitations.
• Dependence on Successful Implementation: The significance of the study relies on the successful implementation of the research. If the research process has flaws or unexpected issues arise, the anticipated significance might not be realized.

About the author

Muhammad Hassan

Researcher, Academic Writer, Web developer

You may also like

Research Paper Title Page – Example and Making...

Data Interpretation – Process, Methods and...

Data Collection – Methods Types and Examples

Data Analysis – Process, Methods and Types

Institutional Review Board – Application Sample...

Research Recommendations – Examples and Writing...

Global site navigation

• Celebrity biographies
• Messages - Wishes - Quotes
• TV-shows and movies
• Fashion and style
• Capital Market
• Celebrities
• Family and Relationships

Local editions

• Legit Nigeria News
• Legit Hausa News
• Legit Spanish News
• Legit French News

Background of the study in research: guide on how to write one

The background of the study is one of the key aspects you need to get right when you are writing a research paper. It is the key to introducing your readers to the topic of your research, and it is different from the lead part. Here is how to write the background of the study in research studies.

Here is all you need to know about the study's science background and how to write one.

What is the background of the study?

Background of the study meaning: The background of the study is a part of the research provided in the introduction section of the paper. It is a compilation of adequate information that is based on the analysis of the problem or proposed argument, the steps and methods needed to arrive at the design, the implementation of the results achieved, and feasible solutions.

It is different from the introduction. The introduction only contains preliminary information about your research question or thesis topic. It is simply an overview of the research question or thesis topic. But the background of your study is more in-depth - it explains why your research topic or thesis is worth your readers' time.

Exclusive: Former presidential aide reveals why PDP will dislodge APC, win presidential election in 2023

The background of the study will provide your readers with context to the information discussed throughout your research paper. It can include both relevant and essential studies.

The background of the study is used to prove that a thesis question is relevant and also to develop the thesis. In summary, a good background of the study is the work done to determine that your research question or thesis topic is a problem and that the method used is the one required to solve the issue or answer the question.

What is the importance of the background of the study?

The background of the study helps your reader determine if you have a basic understanding of the research problem being investigated and promotes confidence in the overall quality of your analysis and findings.

How to write the background of the study in a research paper

• Stage 1. At the beginning stages of formulating your thesis, many of the issues are still very unclear, and you need to solidify your thoughts, so you should conduct preliminary research. This will help you put forward a research question or thesis statement that will lead to more relevant and specific research. You can visit a library, check the internet and other electronic databases to find preliminary sources such as scholarly journals and books about your background of the study.
• Stage 2. Read and gather the info you need to develop - a thesis statement or research question that will guide your thesis. You should take notes and keep an accurate track of the sources of information you have used up to this point. Many people use note cards, but it’s easier and better to use electronic note-taking programs in this electronic age. Just make sure to use a form that is comfortable and easier for you. Also, make sure you cite the source of every piece of information you are using on each note so that you won’t forget where you got the information from, just in case you want to use it in your thesis.
• Stage 3. Develop and pen down the research question or thesis statement. Think about the things you’ve read and searched and the issues or solutions that have been found by other people, and then formulate your stance or opinion on the issue. Write out your position or opinion as an authoritative statement. You may conduct more detailed research at this point and look for more sources that are more relevant to your research question or thesis.
• Stage 4. Complete your research using the thesis statement as your guide. Find sources that are relevant to your specific thesis and provide more insight into your research question using these sources. Your sources should provide information on your thesis's history and past research.
• Stage 5. As you create your background study, create relevant sections. When you start writing, create five sections with the key issues, major findings, and controversies surrounding your thesis, and a section that provides evaluation and a conclusion.
• Stage 6. Identify the further studies that need to be done in the conclusion section. Also, mention possible solutions to the issues that have not been considered in the past.
• Stage 7. Revise and edit your background of the study carefully. You can write out several drafts of your work, revising, editing, and adding more information before coming up with the final one. Make sure each draft is better than the previous one. You can also ask someone else to help you go through it.

Call of the void: why do we feel the urge to jump off high places?

The writing format

You can follow this format when writing your background of the study:

• Start by giving a general overview of your thesis topic and introduce the key ideas you will use throughout your thesis.
• Then, give precise information about all the methodologies used in the research. This can take up to several paragraphs depending on the individual and research question or thesis topic.
• Cite your sources where necessary to avoid plagiarism.
• Then you can briefly introduce the experiment by describing your choice of methodology, why you have decided to use this methodology instead of others, and the objective of the methodology.

What does a good background of the study example contain?

A good example of the background of the study is one that:

• Contains reviews of the area being researched
• Has currently available information about the problem of the study
• Captures the previous studies on the issue
• Indicates the history of the issue of the study from previous research done on the subject

Amazing 5 Forbes Tips for Investors, Entrepreneurs on Building Business, Brand and Ideas in Nigeria

A good background of the study sample has all these qualities.

How is the background of the study different from the literature review?

The section of literature review follows the background of the study section. It is the second section of your thesis. The literature review supports the study section's background by providing evidence for the proposed hypothesis.

Hopefully, this information on the background of the study has been helpful to you. Read other useful posts on our website to improve your writing skills.

READ ALSO: Can you start a sentence with but: A grammatical explanation

Legit.ng reported that one of the most deeply ingrained grammar rules involves the usage of the word 'but'. For a long time, teachers have told their students that they cannot use conjunctions at the beginning of sentences. If you asked your English teacher, "Can you start a sentence with but?" you would be met with a resounding 'No!'

Tianjin Binhai: Photos and video show Chinese big library with 1.2 million books and tall shelves

But where did this 'rule' even come from? And does it hold water? Find out by reading Legit.ng's take on the usage of 'but' at the beginning of a sentence.

Source: Legit.ng

Adrianna Simwa (Lifestyle writer) Adrianna Simwa is a content writer at Legit.ng where she has worked since mid-2022. She has written for many periodicals on a variety of subjects, including news, celebrities, and lifestyle, for more than three years. She has worked for The Hoth, The Standard Group and Triple P Media. Adrianna graduated from Nairobi University with a Bachelor of Fine Arts (BFA) in 2020. In 2023, Simwa finished the AFP course on Digital Investigation Techniques. You can reach her through her email: [email protected]

How to write the background of the study for a research proposal?

The background of the study is an important part of the research proposal or the thesis dissertation.

The gaps you identify will set up the foundations for the next section of your proposal: the problem statement. You can read more about how to write the problem statement here .

Table of Contents

Example of background of the study.

Recently, machine learning methodologies have been widely used for developing IDSs, with Deep Learning as a new trend for building IDSs (Ponkarthika & Saraswathy, 2018; Prajyot & Kalavadekar, 2018; Tang, Mhamdi, McLernon, Zaidi, & Ghogho, 2018; Yin et al., 2017). According to Elsherif (2018), an IDS can be built using a Recurrent Neural Network (RNN) which is a Deep Learning approach. A Recurrent Neural Network-Intrusion Detection System (RNN-IDS) is able to identify seen and unseen threats, as it has a strong modelling ability, provides a highly accurate detection rate, and low false-positive rate, especially when using NSL-KDD dataset (Elsherif, 2018). However, RNN-IDS tend to be biased, as dataset from real-world production networks is mostly not used (Elsherif, 2018) during the training process. Mostly only benchmark datasets are used to train IDSs and the benchmark datasets are not good representatives of real-world traffic (Bhuyan, Bhattacharyya, & Kalita, 2015). This increases the need to use datasets from real-world networks when developing RNN-IDSs, for them not to be biased in the real-world environment in which they will operate. ” (Aludhilu H.N.; An intrusion detection system using a recurrent neural network with a real-world dataset. 2019.)

Whit this background, the researcher is ready to state the problem statement that will be solved with the research.

Background of study vs. Literature Review

This is a common mistake and should avoid it.

On the other hand, in the literature review, you write about how other researchers try to solve (or solved) the problem you stated.

After you write the problem you are planning to solve, then you write the literature review. This section (a chapter if you are writing the dissertation) is all about other solutions related to the same problem you want to solve.

When you write the background of the study for your research proposal you must:

• Writing Resources
• Free Report
• Get a Quote

Tuesday 11 July 2017

Research proposal: motivation and background, motivation for your research.

1.    Introduction

2.    background, 3.    research questions.

Style and Grammar Guidelines

APA Style provides a foundation for effective scholarly communication because it helps writers present their ideas in a clear, concise, and inclusive manner. When style works best, ideas flow logically, sources are credited appropriately, and papers are organized predictably. People are described using language that affirms their worth and dignity. Authors plan for ethical compliance and report critical details of their research protocol to allow readers to evaluate findings and other researchers to potentially replicate the studies. Tables and figures present information in an engaging, readable manner.

The style and grammar guidelines pages present information about APA Style as described in the Publication Manual of the American Psychological Association, Seventh Edition and the Concise Guide to APA Style, Seventh Edition . Any updates to APA Style are noted on the applicable topic pages. If you are still using the sixth edition, helpful resources are available in the sixth edition archive .

Looking for more style?

• Accessibility of APA Style
• Line Spacing
• Order of Pages
• Page Header
• Paragraph Alignment and Indentation
• Sample Papers
• Title Page Setup
• Appropriate Level of Citation
• Basic Principles of Citation
• Classroom or Intranet Sources
• Paraphrasing
• Personal Communications
• Quotations From Research Participants
• Secondary Sources
• Abbreviations
• Capitalization
• Italics and Quotation Marks
• Punctuation
• Spelling and Hyphenation
• General Principles for Reducing Bias
• Historical Context
• Intersectionality
• Participation in Research
• Racial and Ethnic Identity
• Sexual Orientation
• Socioeconomic Status
• Accessible Use of Color in Figures
• Figure Setup
• Sample Figures
• Sample Tables
• Table Setup
• Archival Documents and Collections
• Basic Principles of Reference List Entries
• Database Information in References
• DOIs and URLs
• Elements of Reference List Entries
• Missing Reference Information
• Reference Examples
• References in a Meta-Analysis
• Reference Lists Versus Bibliographies
• Works Included in a Reference List
• Active and Passive Voice
• Anthropomorphism
• First-Person Pronouns
• Logical Comparisons
• Plural Nouns
• Possessive Adjectives
• Possessive Nouns
• Singular “They”
• Adapting a Dissertation or Thesis Into a Journal Article
• Correction Notices
• Cover Letters
• Journal Article Reporting Standards (JARS)
• Open Science
• Response to Reviewers

consultant and research

• Grade 7 paper topic ideas
• Technology paper questions
• Format issues
• Sample APA project title pages
• Common mistakes of writing
• WWI research paper ideas
• Ancient world: term paper ideas
• Term paper ideas on current events
• Paper questions: government
• To Kill A Mockingbird
• Writing ideas for a finance paper
• Biology topic suggestions
• Paper topics on modern computers
• Writing ideas: voting
• Criminology project ideas
• Zoos: great writing prompts
• Malcolm X: research project ideas
• Developmental psychology topics
• Ideas on modern US history
• Psychology paper tips
• Horses: project topics
• Writing ideas on accounting
• Topics about child abuse
• 9/11 research project writing advice
• Search directions
• Free assistance
• Study background examples
• Abstract writing tips
• Getting APA paper samples
• Purchasing a custom term paper
• Finding project samples: Education
• Academic paper writing
• Finding paper editing companies
• Buying term papers with no risks
• Writing a research project outline
• General writing advice
• Example paper proposals
• Completing a history project
• Books and Web: paper tips
• Hints on literature review
• Computer science paper advice
• Visualization paper hints
• Finding a proper writer
• Qualified help with term papers
• Asthma: paper tips
• Choosing a writing company
• In quest of skilled writers for hire
• How to find a proper helper
• A guide to the writer's search
• Why to purchase projects
• Fresh tips on papers for sale
• Writing service selection
• Writing hints

Looking For A Good Research Paper Background Of The Study Sample

As a student, you may be looking for a good research paper background of the study sample. In some cases the student may know nothing about the topic he is going to research so it is good to get some background information. There are numerous things to look for when you are looking for a good research paper background.

• - A definition of the topic you have been assigned
• - A summary of the essence of the topic you have been assigned
• - A list of the names of people who are considered experts in the subject
• - A Time line of major events
• - A discussion of the main issues
• - Specific terminology that can be used to search for information
• - Additional sources derived from bibliographies

If a student has this information, he will have a starting point with which to gather more information. It will give him guidance and to where to find additional information. The summary and definitions will help the writer decide on a thesis statement. He will probably have to go considerable research before he can begin to figure out the thesis statement. Hopefully he will have access to study samples to find out more information. Once he has all of the research, he should have a good understanding of the topic and be able to formulate a thesis statement.

Once he has the topic and the topic statement, he can create his research paper. He has done the sufficient research and hopefully it is organized so he can access the information he needs. Now he has to create an outline so he can guide his way through the article. The outline will be a road map of his paper. It will highlight the important points that he wants to include. This way they won’t be forgotten. The paper will flow much better if there is an outline to keep you focused and on track.

After the outline is completed, it is time to create your rough draft. This is easy now that you have all the information and the order in which you are going to present it. Once your rough draft is complete, show it to your professor for some guidance in any changes. After that, you can complete your final draft. It is absolutely crucial for you to be looking for research paper background so you can familiarize yourself with your topic. Your study sample will be adequately examined and a great paper will be the result.

Hire a professional paper writing service to complete your term paper or essay in few hours or days.

Writing Help

ideas to try right now.

• Buying research projects with no risk
• College term project tips: bibliography
• Working with project writing companies
• General paper writing advice
• Essay writers online
• Project ideas on natural disasters
• Excellent term project writing advice
• Inventing a Psychology paper question
• Topics for a middle school paper
• Secure methods to buy projects
• A guide to the research project structure

• How to Order

Research Paper Guide

Research Paper Example

Research Paper Examples - Free Sample Papers for Different Formats!

People also read

Research Paper Writing - A Step by Step Guide

Guide to Creating Effective Research Paper Outline

Interesting Research Paper Topics for 2024

Research Proposal Writing - A Step-by-Step Guide

How to Start a Research Paper - 7 Easy Steps

How to Write an Abstract for a Research Paper - A Step by Step Guide

Writing a Literature Review For a Research Paper - A Comprehensive Guide

Qualitative Research - Methods, Types, and Examples

8 Types of Qualitative Research - Overview & Examples

Qualitative vs Quantitative Research - Learning the Basics

200+ Engaging Psychology Research Paper Topics for Students in 2024

Learn How to Write a Hypothesis in a Research Paper: Examples and Tips!

20+ Types of Research With Examples - A Detailed Guide

Understanding Quantitative Research - Types & Data Collection Techniques

230+ Sociology Research Topics & Ideas for Students

How to Cite a Research Paper - A Complete Guide

Excellent History Research Paper Topics- 300+ Ideas

A Guide on Writing the Method Section of a Research Paper - Examples & Tips

How To Write an Introduction Paragraph For a Research Paper: Learn with Examples

Crafting a Winning Research Paper Title: A Complete Guide

Writing a Research Paper Conclusion - Step-by-Step Guide

Writing a Thesis For a Research Paper - A Comprehensive Guide

How To Write A Discussion For A Research Paper | Examples & Tips

How To Write The Results Section of A Research Paper | Steps & Examples

Writing a Problem Statement for a Research Paper - A Comprehensive Guide

Finding Sources For a Research Paper: A Complete Guide

A Guide on How to Edit a Research Paper

200+ Ethical Research Paper Topics to Begin With (2024)

300+ Controversial Research Paper Topics & Ideas - 2024 Edition

150+ Argumentative Research Paper Topics For You - 2024

How to Write a Research Methodology for a Research Paper

Crafting a comprehensive research paper can be daunting. Understanding diverse citation styles and various subject areas presents a challenge for many.

Without clear examples, students often feel lost and overwhelmed, unsure of how to start or which style fits their subject.

Explore our collection of expertly written research paper examples. We’ve covered various citation styles and a diverse range of subjects.

So, read on!

• 1. Research Paper Example for Different Formats
• 2. Examples for Different Research Paper Parts
• 3. Research Paper Examples for Different Fields
• 4. Research Paper Example Outline

Research Paper Example for Different Formats

Following a specific formatting style is essential while writing a research paper . Knowing the conventions and guidelines for each format can help you in creating a perfect paper. Here we have gathered examples of research paper for most commonly applied citation styles :

Social Media and Social Media Marketing: A Literature Review

APA Research Paper Example

APA (American Psychological Association) style is commonly used in social sciences, psychology, and education. This format is recognized for its clear and concise writing, emphasis on proper citations, and orderly presentation of ideas.

Here are some research paper examples in APA style:

Research Paper Example APA 7th Edition

Research Paper Example MLA

MLA (Modern Language Association) style is frequently employed in humanities disciplines, including literature, languages, and cultural studies. An MLA research paper might explore literature analysis, linguistic studies, or historical research within the humanities. 

Here is an example:

Found Voices: Carl Sagan

Research Paper Example Chicago

Chicago style is utilized in various fields like history, arts, and social sciences. Research papers in Chicago style could delve into historical events, artistic analyses, or social science inquiries. 

Here is a research paper formatted in Chicago style:

Chicago Research Paper Sample

Research Paper Example Harvard

Harvard style is widely used in business, management, and some social sciences. Research papers in Harvard style might address business strategies, case studies, or social policies.

View this sample Harvard style paper here:

Harvard Research Paper Sample

Examples for Different Research Paper Parts

A research paper has different parts. Each part is important for the overall success of the paper. Chapters in a research paper must be written correctly, using a certain format and structure.

The following are examples of how different sections of the research paper can be written.

Research Proposal

The research proposal acts as a detailed plan or roadmap for your study, outlining the focus of your research and its significance. It's essential as it not only guides your research but also persuades others about the value of your study.

Example of Research Proposal

An abstract serves as a concise overview of your entire research paper. It provides a quick insight into the main elements of your study. It summarizes your research's purpose, methods, findings, and conclusions in a brief format.

Research Paper Example Abstract

Literature Review 

A literature review summarizes the existing research on your study's topic, showcasing what has already been explored. This section adds credibility to your own research by analyzing and summarizing prior studies related to your topic.

Literature Review Research Paper Example

Methodology

The methodology section functions as a detailed explanation of how you conducted your research. This part covers the tools, techniques, and steps used to collect and analyze data for your study.

Methods Section of Research Paper Example

How to Write the Methods Section of a Research Paper

The conclusion summarizes your findings, their significance and the impact of your research. This section outlines the key takeaways and the broader implications of your study's results.

Research Paper Conclusion Example

Research Paper Examples for Different Fields

Research papers can be about any subject that needs a detailed study. The following examples show research papers for different subjects.

History Research Paper Sample

Preparing a history research paper involves investigating and presenting information about past events. This may include exploring perspectives, analyzing sources, and constructing a narrative that explains the significance of historical events.

View this history research paper sample:

Many Faces of Generalissimo Fransisco Franco

Sociology Research Paper Sample

In sociology research, statistics and data are harnessed to explore societal issues within a particular region or group. These findings are thoroughly analyzed to gain an understanding of the structure and dynamics present within these communities. 

Here is a sample:

A Descriptive Statistical Analysis within the State of Virginia

Science Fair Research Paper Sample

A science research paper involves explaining a scientific experiment or project. It includes outlining the purpose, procedures, observations, and results of the experiment in a clear, logical manner.

Here are some examples:

Science Fair Paper Format

What Do I Need To Do For The Science Fair?

Psychology Research Paper Sample

Writing a psychology research paper involves studying human behavior and mental processes. This process includes conducting experiments, gathering data, and analyzing results to understand the human mind, emotions, and behavior.

Here is an example psychology paper:

The Effects of Food Deprivation on Concentration and Perseverance

Art History Research Paper Sample

Studying art history includes examining artworks, understanding their historical context, and learning about the artists. This helps analyze and interpret how art has evolved over various periods and regions.

Check out this sample paper analyzing European art and impacts:

European Art History: A Primer

Research Paper Example Outline

Before you plan on writing a well-researched paper, make a rough draft. An outline can be a great help when it comes to organizing vast amounts of research material for your paper.

Here is an outline of a research paper example:

Here is a downloadable sample of a standard research paper outline:

Research Paper Outline

Want to create the perfect outline for your paper? Check out this in-depth guide on creating a research paper outline for a structured paper!

Good Research Paper Examples for Students

Here are some more samples of research paper for students to learn from:

Fiscal Research Center - Action Plan

Qualitative Research Paper Example

Research Paper Example Introduction

How to Write a Research Paper Example

Research Paper Example for High School

Now that you have explored the research paper examples, you can start working on your research project. Hopefully, these examples will help you understand the writing process for a research paper.

If you're facing challenges with your writing requirements, you can hire our essay writing help online.

Our team is experienced in delivering perfectly formatted, 100% original research papers. So, whether you need help with a part of research or an entire paper, our experts are here to deliver.

So, why miss out? Place your ‘ write my research paper ’ request today and get a top-quality research paper!

Write Essay Within 60 Seconds!

Nova Allison is a Digital Content Strategist with over eight years of experience. Nova has also worked as a technical and scientific writer. She is majorly involved in developing and reviewing online content plans that engage and resonate with audiences. Nova has a passion for writing that engages and informs her readers.

Paper Due? Why Suffer? That’s our Job!

Keep reading

Research Paper Topic Approval Form Example

Poster Samples

Looking at samples of real student posters can help you generate ideas and define your goals. As you get started, it may be helpful to look at examples of finished posters.

Below are a number of sample posters created by UT undergraduates. There is a brief discussion of each poster highlighting its greatest strengths and areas where there is room for improvement.

• More than one type of visual aid
• Logical order for sections
• Acknowledgments

Room for improvement

• Background may be distracting, or detract from content
• Sections and images are not aligned
• Too many visual components clutter poster

• White space
• Legible text and graphics
• Reports preliminary results
• All participants listed as authors, with affiliations provided
• Lacks Citations and Acknowledgements
• Labeling of images/graphics
• Inconsistent text alignment
• Color-saturated background

• Clearly defined research questions
• Effective use of visual aids
• Clear organizational structure
• Bullets break up text
• Technical language/undefined acronyms (accessible to limited audience)
• Narrow margins within text boxes
• Too many thick borders around boxes
• Uses UT seal instead of college or university wordmark

• Clear introductory material
• Use of bullet points
• Logical flow
• Color-coding in graphics
• Lacks references section
• May not be accessible to all audiences (some technical language)
• No need for borders around sections (the blue headers are sufficient)

• Compelling visual aids
• Strategic use of color
• Clear sections
• Inconsistent fonts in body text
• Abstract section mislabeled
• Bullet points are great, but only if they’re used judiciously

• Parameters of study well defined
• Clearly defined research question
• Simple color scheme
• Use of white space
• Discussion of Results
• Minor formatting misalignments
• Unauthorized use of UT seal (use wordmark instead)

• Venn diagram in discussion
• Consistent graphics
• Multiple types of visual aids
• Light text on dark background
• Color backgrounds should be avoided, especially dark ones
• Unlabeled, non-credited photos

• Easy to read
• Use of shapes, figures, and bullets to break up text
• Compelling title (and title font size)
• Clean overall visual impression
• Many sections without a clear flow between them
• Lacks acknowledgements

• Use of images/graphics
• Clear title
• Accessible but professional tone
• Length/density of text blocks
• Tiny photo citations
• Connections between images and descriptive text
• Vertical boxes unnecessary

• Compelling title
• Font sizes throughout (hierarchy of text)
• Simple graphics
• Lacks clear Background section
• Relationship of Findings and Conclusion to Research questions

• Use of visual aids
• Uneven column width
• Center-justfied body text
• Lacks “Methods” section

• Use of bullets
• Too many different font styles (serif and sans serif, bold and normal)
• Concise interpretation of graphics

• Accessible visual structure
• Clear, simple graphics
• Fonts and font sizes
• Analysis of graphic data
• Discussion of significance
• Lacks author’s affiliation and contact information

• Balance among visuals, text and white space
• Data presented in visual format (SmartArt)
• Accesible to many audiences (simple enough for general audience, but enough methodological detail for experts)
• Some more editing needed
• When targeting an expert audience (as in the methodology section), should also report statistics ( r, p, t, F, etc.)

• Large, clear title
• Creative adaptation of sections
• Use of lists (rather than paragraphs)
• Accessible to diverse audience
• Connection between visuals (sheet music) and content

• Strategic use of color for section headers
• Labeling and citation of images
• Accessible to a broad audience
• Wide margins around poster edges
• Slightly text-heavy
• Data referenced (“Methodology”) but not discussed

What is my next step?

Begin working on the content for your poster at Create Your Message .

This paper is in the following e-collection/theme issue:

Published on 26.8.2024 in Vol 26 (2024)

Sociotechnical Cross-Country Analysis of Contextual Factors That Impact Patients’ Access to Electronic Health Records in 4 European Countries: Framework Evaluation Study

Authors of this article:

Original Paper

• Jonas Moll 1 , PhD   ; 
• Isabella Scandurra 1 † , PhD   ; 
• Annika Bärkås 2, 3 , PhD   ; 
• Charlotte Blease 2, 4 , PhD   ; 
• Maria Hägglund 2, 3 , PhD   ; 
• Iiris Hörhammer 5 , DSC   ; 
• Bridget Kane 2, 6 , PhD   ; 
• Eli Kristiansen 7 , ME   ; 
• Peeter Ross 8, 9 , MD, PhD   ; 
• Rose-Mharie Åhlfeldt 10 , PhD   ; 
• Gunnar O Klein 1 , MD, PhD  

1 Centre for Empirical Research on Information systems, School of Business, Örebro University, Örebro, Sweden

2 Participatory eHealth and Health Data Research Group, Department of Women's and Children's Health, Uppsala University, Uppsala, Sweden

3 Medtech Science & Innovation Centre, Uppsala University Hospital, Uppsala, Sweden

4 Digital Psychiatry, Department of Psychiatry Beth Israel Deaconess Medical Center, Harvard Medical School, Boston, MA, United States

5 Department of Computer Science, Aalto University, Espoo, Finland

6 Karlstad University Business School, Karlstad, Sweden

7 Norwegian Centre for E-health Research, University Hospital of North Norway, Tromsø, Norway

8 E-Medicine Centre, Department of Health Technologies, Tallinn University of Technology, Tallinn, Estonia

9 Research Department, East Tallinn Central Hospital, Tallinn, Estonia

10 School of Informatics, University of Skövde, Skövde, Sweden

Corresponding Author:

Jonas Moll, PhD

Centre for Empirical Research on Information systems

School of Business

Örebro University

Nova building, 4th floor

Fakultetsgatan 1

Örebro, 70182

Phone: 46 0703174383

Email: [email protected]

Background: The NORDeHEALTH project studies patient-accessible electronic health records (PAEHRs) in Estonia, Finland, Norway, and Sweden. Such country comparisons require an analysis of the sociotechnical context of these services. Although sociotechnical analyses of PAEHR services have been carried out in the past, a framework specifically tailored to in-depth cross-country analysis has not been developed.

Objective: This study aims to develop and evaluate a method for a sociotechnical analysis of PAEHRs that advances a framework for sociotechnical analysis of eHealth solutions first presented by Sittig and Singh. This first article in a series presents the development of the method and a cross-country comparison of the contextual factors that enable PAEHR access and use.

Methods: The dimensions of the framework for sociotechnical analysis were thoroughly discussed and extended in a series of workshops with international stakeholders, all being eHealth researchers focusing on PAEHRs. All countries were represented in the working group to make sure that important national perspectives were covered. A spreadsheet with relevant questions related to the studied services and the various dimensions of the sociotechnical framework was constructed and distributed to the 4 participating countries, and the project participants researched various national sources to provide the relevant data for the comparisons in the 10 sociotechnical dimensions.

Results: In total, 3 dimensions were added to the methodology of Sittig and Singh to separate clinical content from features and functions of PAEHRs and demonstrate basic characteristics of the different countries regarding national and regional steering of health care and information and communications technology developments. The final framework contained the following dimensions: metadata; hardware and software computing infrastructure; features and functions; clinical content shared with patients; human-computer interface; people; workflow and communication; the health care organization’s internal policies, procedures, and culture; national rules, regulations, and incentives; system measurement and monitoring; and health care system context. The dimensions added during the study mostly concerned background information needed for cross-country comparisons in particular. Several similarities were identified among the compared countries, especially regarding hardware and software computing infrastructure. All countries had, for example, one national access point, and patients are provided a PAEHR automatically. Most of the differences could be identified in the health care system context dimension. One important difference concerned the governing of information and communications technology development, where different levels (state, region, and municipality) were responsible in different countries.

Conclusions: This is the first large-scale international sociotechnical analysis of services for patients to access their electronic health records; this study compared services in Estonia, Finland, Norway, and Sweden. A methodology for such an analysis was developed and is presented to enable comparison studies in other national contexts to enable future implementations and evaluations of PAEHRs.

Introduction

Patients’ web-based access to their electronic health records (EHRs) is increasingly being implemented internationally [ 1 - 3 ], and a growing body of literature indicates strong benefits to patients, including understanding their care plans better [ 4 ], feeling more in control of their care [ 4 , 5 ], being better informed about their medication [ 6 ], improved communication with and trust in their clinicians [ 5 , 7 ], and improved patient safety [ 8 ].

Patient-accessible EHRs (PAEHRs) are web-based services providing patients with secure access to view and sometimes edit or comment on their EHRs made available by their health care providers [ 9 ]. A PAEHR is directly linked to the EHRs, which are shared patient records entered and maintained by health care service providers and contain historical data about a patient [ 10 ]. Medical and health data and information in the EHR are created and managed by authorized providers in a digital format capable of being shared with other providers across more than one health care organization [ 11 ]. In addition, a PAEHR may include access to the services supporting a person’s access to health care services (eg, e-booking and e-consultation) and to evidence-based tools (eg, patient guidelines, educational materials, and reimbursement information). It is important to distinguish between PAEHRs and personal health records—the latter being a health record that the patients themselves control and maintain to track their own health, for example, on paper; in Microsoft Excel sheets; or, predominantly, in health care apps. Personal health records are external applications that can potentially be linked with EHRs to enable sharing [ 10 ].

In this study, we focused on national services in the 4 countries of Estonia, Finland, Norway, and Sweden and excluded web-based services designed to provide access to the services of only 1 health care provider. Among the studied countries, Estonia was the first to offer citizens web-based access to their EHR through the service Digilugu that was launched nationwide in 2008 [ 12 ]. The Finnish counterpart, Omakanta, was launched nationwide in 2010 with only limited functionality and reached full functionality in 2015 after a step-by-step adoption of functions [ 13 ]. In Sweden, the national PAEHR service Journalen was launched in the Region Uppsala in 2012 and has been accessible to all citizens since 2018 [ 14 ]. In Norway, the national PAEHR service Helsenorge was first launched in 1 of 4 regions in 2015, and as of 2023, citizens in 3 out of 4 regions can use the service to access their PAEHRs [ 15 ].

Despite the reported benefits of PAEHRs and patients’ web-based record access, implementation is often slow and challenging, and the complexity of health care systems and technical infrastructure leads to great diversity in, for example, the information to which patients are given access and when they can access it across regions and health care settings. Similarly, patients’ adoption and use of PAEHRs also varies across contexts. To understand why these differences exist and better adapt the design and implementation of PAEHRs to a specific context, there is a need for a more fine-grained understanding of the social and technical underpinnings of this innovation.

Sociotechnical Systems

As described by Baxter and Sommerville [ 16 ], the problems that arise when designing and implementing complex IT systems are not just technical, engineering problems. These systems are developed and operated by people working in organizations that inevitably have different, often conflicting goals and views on the role and design of the system. The IT system is part of a broader “sociotechnical” system, and to understand success factors and barriers to implementing web-based record access and identify best practices and guidelines, there is a need to approach these eHealth services as complex sociotechnical systems.

Sittig and Singh [ 17 ] have proposed a multidimensional sociotechnical framework in which any health IT innovation, intervention, application, or device implemented within a complex adaptive health care system can be studied. The sociotechnical framework by Sittig and Singh [ 17 ] identifies eight dimensions of sociotechnical systems in health care that need to be considered in both development and evaluation: (1) hardware and software computing infrastructure; (2) clinical content; (3) human-computer interface; (4) people; (5) workflow and communication; (6) internal organizational policies, procedures, and culture; (7) external rules, regulations, and pressures; and (8) system measurement and monitoring [ 17 ]. The framework breaks down components of the technology to enable researchers to identify specific problems with implementation. It also includes monitoring processes and government structures that need to be in place for the system to achieve its goals. The interrelatedness of the components makes the framework pertinent when eHealth technologies and users are at the core of the investigation.

In 2017, Hägglund and Scandurra [ 18 ] began analyzing the Swedish PAEHR system from a sociotechnical perspective using the framework by Sittig and Singh [ 17 ], and their results laid the foundation for continued work within the NORDeHEALTH research project [ 19 ] involving partners from Estonia, Finland, Norway, and Sweden. These countries were deemed appropriate for the development and first test of the cross-country sociotechnical analysis method derived in this study as they have comparable government and health care structures. In addition, they have all reached maturity in the use of their respective PAEHR systems—these countries are all among the early adopters of this specific type of eHealth service for patients. In a 2021 workshop, a first draft of an extended version of the framework by Sittig and Singh [ 17 ] as a sociotechnical template for analyzing PAEHR implementations was introduced and discussed [ 20 ]. However, the original framework by Sittig and Singh [ 17 ] needed to be more specifically adapted to PAEHRs and their contexts to enable useful cross-country comparisons, a process that will be described further in this paper.

The full results of the sociotechnical analysis of the 4 countries involved in the NORDeHEALTH project are being published in an article series; it is intended for each publication to offer results related to specific and related framework dimensions and, thereby, focus on different themes. This paper is intended to offer a detailed description of the overarching research method as well as a sociotechnical analysis of dimensions providing the context for other articles of the series.

Aims and Research Questions

The aims of this study were to (1) develop a detailed collaborative method suitable for cross-country sociotechnical analyses of PAEHRs and (2) use the data collection to compare contextual factors that enable PAEHR access and use.

Although the proposed work process, including the production of an extensive sociotechnical analysis template, is described in the context of comparing PAEHRs in 4 specific countries, it can be adjusted to cover other eHealth services for patients across other nations. The comparison and discussion regarding the national contexts will provide a robust foundation and understanding of other parts of the sociotechnical analysis, such as included features and national and regional incentives for the use and promotion of PAEHRs. The following research questions (RQs) guided the work presented in this paper:

• RQ 1: how can a sociotechnical framework for health IT be adapted and a collaborative method be developed that is suitable for cross-country sociotechnical analyses of PAEHR services?
• RQ 2: how can the method be used to compare the PAEHRs in Estonia, Finland, Norway, and Sweden?

Research Method

This study has 2 main parts inspired by different phases in the design science methodology [ 21 ]. In the first part, a collaborative method, including a data collection template, for sociotechnical cross-country analysis of PAEHRs was derived: (1) problem identification and motivation; that is, based on earlier experiences with the framework by Sittig and Singh [ 17 ] and the fact that the framework is not adapted for comparisons across contexts, the following problem was defined, how can we conduct fruitful comparative analysis among PAEHRs in different regions and health care settings? (2) objectives of a solution; that is, due to the identified need to perform a complete sociotechnical comparison among health care contexts, the main objective of the new artifact (being a method for collecting data, including a refined framework and a data collection instrument) is to enable a comparison that would consider the sociotechnical contexts of PAEHRs; and (3) design and development; that is, a dedicated collaborative method was developed with a focus on adjusting the framework for cross-country comparisons and deriving a data collection template enabling the collection of all relevant data from the specific countries involved.

The second part of our study (in this paper referred to as data collection ) consists of the collection and comparison of the sociotechnical characteristics of the 4 countries participating in the NORDeHEALTH project: Estonia, Finland, Norway, and Sweden. This can be seen as case studies of the more general questions regarding the availability and use of the PAEHR in a country. However, it is also an evaluation of the method as an artifact in the design science paradigm: (1) demonstration—the data collection template derived from the first part of the study was used to conduct a cross-country comparison among 4 Nordic countries—and (2) evaluation—finally, the results from the comparison were analyzed to evaluate whether the approach provides an understanding that explains differences among countries in the adoption of PAEHRs by regions and providers and patients.

In the remaining sections, the different points of the process are elaborated on, with emphasis on the design and development, demonstration, and evaluation steps. The work was led by a core analysis team (JM, IS, GK, and AB) of health informatics researchers.

Development of the Collaborative Method

The development of the data collection instrument based on the sociotechnical framework proposed by Sittig and Singh [ 17 ] started before the NORDeHEALTH project had begun.

In 2017, a first sociotechnical analysis of the Swedish PAEHR service Journalen was conducted with the aim of increasing the understanding of factors that influence the design, implementation, adoption, and use of the service [ 18 ]. However, the analysis did not go into the details of each dimension of the framework but, rather, highlighted some overall challenges. The results of this early analysis were used as input to an international workshop held in 2021, for which a more detailed template specifically adapted to PAEHRs was developed ( Textbox 1 ). In addition to including specific PAEHR-related questions to each dimension, a new dimension, Features and functions , was added. Some other dimensions were renamed. During the workshop, international experts provided feedback on the template that was incorporated into further refinements, leading to the data collection form presented in this study.

Dimension and description

• Hardware and software computing infrastructure: focuses only on the hardware and software required to run the applications.
• Features and functions: important features and functions in the patient-accessible electronic health record (PAEHR) service or in related services. As there is not yet a strict definition of what functionality is included in a PAEHR or not, we included functions that may be considered external to a PAEHR in some contexts.
• Clinical content shared with patients: includes everything on the data-information-knowledge continuum that is stored in the PAEHR service and made accessible to patients.
• Human-computer interface: focuses on the usability of the PAEHR service.
• People: represents the humans involved in all aspects of the implementation and use of the eHealth application and how they experience the use.
• Workflow and communication: focuses on collaboration and communication among different users and assessing how well the eHealth application supports the current clinical workflow.
• The health care organization’s internal policies, procedures, and culture: affects every other dimension in this model as it includes any internal IT policy documents and managerial procedures that may influence the implementation and use of eHealth.
• National rules, regulations, and incentives: focuses on external forces that facilitate or place constraints on the design, development, implementation, use, and evaluation of eHealth in the respective clinical settings.
• System measurement and monitoring: focuses on the need for an effective system measurement and monitoring program to identify the availability of features and functions and how they are used as well as expected outcomes and unintended consequences of the PAEHR service.

The resulting Microsoft Excel template was then used as a basis for developing a complete data collection form within the scope of the NORDeHEALTH project, including questions, response options, and comment sections for all the dimensions in the sociotechnical analysis framework used. The first version of the form was developed by the core analysis team. The sociotechnical dimensions were not changed by the team, but the added “Features and functions” dimension was retained, and the questions used in the Microsoft Excel template were clarified in many cases. The work process followed a weekly workshop format in which different framework dimensions were in focus and that lasted between February 2021 and April 2021. The draft template took form within a shared Google document, enabling collaborators in the participating countries to continuously offer feedback on the ongoing work. This format ensured that the items in the data collection form were relevant in the context of all involved countries. Finally, an analysis dimension called metadata was added during this iterative process, whereby information about the data collection itself—including the name of the system, the name of the researcher responsible for data collection, and information sources—could be noted.

After the first complete draft had been developed, a digital workshop was held in early May 2021 in which the core analysis team from Sweden (n=4) as well as representatives from Estonia (n=1), Finland (n=1), and Norway (n=2) participated. All participants are coauthors of this paper and are health informatics researchers with several years of experience following the implementation and subsequent use of the PAEHR system in their respective countries. The focus of the workshop was to discuss how to interpret the framework dimensions and uppermost to elaborate on and resolve some of the question formulations that had elicited many comments from other project partners in the shared document. After the 1-hour workshop, the template was refined, after which input was again sought from other project partners through email. After the last questions had been resolved, in October 2021, data collection could start based on the finalized template.

Data Collection

Data collection was undertaken in 2 different phases. The first, longer phase started in October 2021 and ended in November 2021. One representative from each country, who also took part in the workshop where the data collection template was discussed, was assigned to be the main responsible data collector from their country, and communication among these representatives occurred continuously during the data collection period. Project participants from each country filled out a copy of the data collection form ( Multimedia Appendix 1 ) for their main PAEHR systems and with the systems shown in Table 1 in focus.

Some ambiguities were identified during the data collection, and these were handled through communication among the responsible researchers. When needed, the data collection master file was updated in accordance with agreed-upon solutions while making sure that everyone always used the question formulations in the latest version. After completion, all main responsible researchers from each country shared their filled-out forms in a shared folder for everyone to access.

After a preliminary walk-through of the collected data performed by the core analysis team, a need for further clarification was identified. In some cases, a few questions remained unanswered, and in other cases, the level of detail varied, requiring follow-up questions. When questions that the researchers from each country needed to elaborate on had been compiled, an additional and final data collection took place in December 2021. This shorter phase of data collection took place through email exchange between the leader of the core analysis team (JM) and the researchers responsible for data collection in each country. Multimedia Appendix 2 summarizes the sources of information that were used during the data collection in each country.

As a last step of data collection for this particular paper, a project representative from each involved country was tasked with collecting and summarizing information to enable a broader overview of national contexts than the developed data collection form could provide. In this final stage, information related to government structure, overall health care system, digital care organization, and steering of health information and communications technology (ICT) developments was gathered from each country. Information was gathered from national statistics, agency web pages, and local contacts within each health care system. Data collection for this study ended in November 2022.

When data collection was completed, the core analysis team began the analysis work. The first step was to copy all answers from the completed data collection forms into a shared analysis document. In this document, each dimension had its own sheet to simplify comparisons within each dimension. In each sheet, the questions from the data collection form were added as rows, and each PAEHR system was added as a column to create a matrix where each column included the answers related to a specific PAEHR system. Each answer, possibly in combination with an additional comment, was added to the corresponding cells in the matrix. Some examples are shown in the Results section. The content of the sheets was then compared across columns (countries) for all questions (rows) to identify similarities as well as aspects that are unique to the specific countries involved. In this step, representatives from Estonia, Finland, Norway, and Sweden were once again invited to discuss the identified similarities and differences.

Ethical Considerations

The study presented in this paper is part of the larger NORDeHEALTH project, which received ethics approval from the appropriate national ethical bodies. This particular study did not involve human participants or sensitive personal data and only focused on contextual and technical details of PAEHRs; there were no specific ethical requirements that needed to be addressed.

Final Data Collection Instrument

The final data collection form inspired by the sociotechnical dimensions in the framework developed by Sittig and Singh [ 17 ], including all updates made during the initial data collection round, can be found in Multimedia Appendix 1 . Table 2 describes the overall content and structure of the form. The data collection form is the end result of the method development and is intended to enable sociotechnical comparisons of eHealth services for patients across countries, with a specific aim to compare PAEHR systems. It is important to note here that the form derived for this study is specifically aimed at comparison of PAEHR systems. For other types of eHealth services for patients, such as self-tracking applications, a different form may have to be derived through the process suggested in the Methods section.

a The questions refer to the numbering in Multimedia Appendix 1 .

b PAEHR: patient-accessible electronic health record.

c EU: European Union.

During the work process, some of the dimensions by Sittig and Singh [ 17 ] were adjusted, and some dimensions were added in response to identified needs that were not fulfilled by the original framework. The following dimensions were either updated or added ( Table 2 ):

• Metadata: this dimension was added during the work process to enable collection of information related to the data collection process itself. It stores general information that is important for keeping track, especially in large-scale projects.
• Features and functions: we decided to use the existing framework dimension Clinical content (after renaming it—see the following item on this list) to describe the information to which patients or citizens are given access through the PAEHR (eg, laboratory test results, medications, and clinical notes). However, there is also variation in the functions that PAEHRs provide (eg, whether patients can comment or fill out forms). Considering the importance of these types of differences, we determined that a new dimension was warranted. Thus, this dimension ensured that the focus was not only on the clinical information that patients have access to in the PAEHR but also on the other important functions or features that they can use.
• Clinical content shared with patients: in the original framework, this dimension was called “Clinical content.” The dimension was renamed to make it more PAEHR specific.
• The health care organization’s internal policies, procedures, and culture: in the original framework, this dimension was more generally about policies, procedures, and culture. The decision was made to add “Health care organizations” as a specification of this dimension. This made it better suited for PAEHR analysis.
• National rules, regulations, and incentives: in the original framework, this dimension was named “External rules, regulations, and pressures.” The redefined dimension puts the focus more on the national context in relation to rules and regulations, again to make it more suitable for PAEHR analysis.
• Health care system context: this dimension was also added during the work process to understand the basics of the countries involved in the comparison. This understanding is needed when analyzing the data in the other dimensions as the national context, including health care system financing and steering of health ICT system development, has an effect on all the other dimensions. It is especially important to gather general data about the national context when comparing across several countries. This dimension was added to the final data collection form after the information had been collected in each country, and hence, it was added in response to a need that the data collection form used did not fulfill.

Identified Clusters of Dimensions

As the collected material is extensive, the dimensions were clustered to enable more manageable partitions of the data set. Dimensions that had similar focuses were grouped together. Textbox 2 presents the 4 resulting clusters. The Contextual factors enabling PAEHR access and use cluster gathers dimensions focusing on the user base as well as technical and governmental prerequisites for PAEHR access and use. The added Metadata dimension is also included in this cluster as it is a prerequisite for the type of comparative studies that we have conducted, and it also includes contextual information. The Features and content cluster includes dimensions focusing on what is actually offered to patients in the PAEHR, that is, the content (eg, clinical notes, test results, and images) and functions (eg, secure messaging and prescription renewal) that are provided to patients. The Evaluations of human-computer interaction and use cluster includes the dimensions focusing on how one measures and evaluates PAEHR use, as well as the results of such evaluations. Both internal service provider evaluations and external evaluations are included to provide a broad coverage. Finally, the National and local policies, regulations for use, promotion, workflow, and communication cluster includes dimensions focusing on laws and regulations; the focus in this cluster is also more on health care professionals and their relationship to PAEHRs than on the patients.

Contextual factors enabling patient-accessible electronic health record access and use

• Hardware and software computing infrastructure
• Health care system context

Features and content

• Features and functions
• Clinical content shared with patients

Evaluations of human-computer interaction and use

• Human-computer interface
• System measurement and monitoring

National and local policies, regulations for use, promotion, workflow, and communication

• Workflow and communication
• The health care organization’s internal policies, procedures, and culture
• National rules, regulations, and incentives

In the article series about the results of the sociotechnical analysis, each article focuses on different clusters from Textbox 2 . This paper has the Contextual factors enabling PAEHR access and use cluster in focus; hence, the dimensions belonging to this cluster were considered in detail, and the results of the data collection in relation to those dimensions are presented in the following sections.

Results From the Sociotechnical Comparison of the “Contextual Factors Enabling PAEHR Access and Use” Cluster

In this section, the results gathered by means of the derived sociotechnical data collection form, with a focus on contextual factors, are presented.

The metadata dimension, which was not part of the original sociotechnical framework presented by Sittig and Singh [ 17 ], includes basic information related to the data collection process itself and the system in question. Collected information about the PAEHR systems is presented in Table 3 . Differences can be observed regarding the type of provider responsible for sharing EHR services in each country. In Sweden, the responsible provider is a publicly owned company, Inera AB, whereas the responsible providers in the other countries are institutions or ministries. In Sweden, Finland, and Estonia, the PAEHR service is national; however, in Norway, the studied service, Helsenorge, is only used for PAEHRs in 3 of 4 regions. All countries only provide 1 national PAEHR service ( Table 3 ); however, local PAEHRs can exist in parallel with the national PAEHR. In all studied countries, health information from different sources is collected in a single PAEHR.

a The items shown in the table correspond to questions 1.1 to 1.4 in the data collection form ( Multimedia Appendix 1 ). Question 1.6 regarding information sources is answered in Multimedia Appendix 2 .

c A total of 2 different Swedish systems (1177 and Journalen) were analyzed. As 1177 includes Journalen (which presents most of the PAEHR information) as well as some other related features, the results from the 2 systems are merged here. In cases in which, for example, the setup of Journalen differs from that of the rest of 1177, the differences will be highlighted.

People and Demographics

This dimension was more complex than the others covered in this paper as demographic data were not available from the same time intervals and age intervals in the different countries. In addition, some of the countries did not have user group statistics of the PAEHR available. Hence, a complete comparison across all countries cannot be made, and as a consequence, results will only be summarized at an overall level in this section.

Sweden has the highest number of inhabitants (10.2 million) of the countries compared, and Estonia has the smallest (1.3 million). Norway and Finland have 5.4 and 5.5 million inhabitants, respectively. Thus, there are large differences in population size among the countries. In addition, there are large differences regarding the proportion of immigrants in the countries, with the lowest proportion (9%) in Finland and the highest proportion (31% non-Estonians) in Estonia. In Norway and Sweden, the proportion is approximately 20%. Data on internet use among the populations of the 4 countries were available from 2019 in Finland and 2020 in Estonia, Norway, and Sweden. The data showed that a high proportion of the populations were internet users, with the highest number in Norway (98%) and the lowest in Estonia (89%). Data from all countries showed a general trend of increasing internet use. Estonia and Norway lack statistics on the use of PAEHRs in different user groups (such as age, gender, and profession). In Finland, statistics on PAEHR use in different age groups are collected, with the highest number of users being aged 36 to 50 years (94%) and 18 to 35 years (93%). In Sweden, data are collected for the PAEHR service Journalen when it comes to age intervals and gender. The service is used most frequently by individuals aged 20 to 29 years and 30 to 39 years, and slightly more female (53%) than male (47%) individuals use it. Educational levels are comparable across countries, with the vast majority of the population reaching upper secondary education. The highest proportion with at least 3 years of higher education can be found in Sweden (37%), and the lowest proportion can be found in Finland and Estonia (23%). No statistics from any of the countries could be found regarding the proportion of the population that prefers an interpreter in their contacts with health care.

Hardware and Software Computing Infrastructure

This dimension includes several important contextual factors that are presented in Table 4 . Among the compared countries, 2 different ways of storing data are represented. In Norway, Finland, and Sweden, data are stored in local EHRs and then extracted and presented in the PAEHRs at runtime. For Digilugu in Estonia, centralized storage is used. In most of the studied countries, data are provided to the PAEHR from both private and public providers and from both primary and secondary care. The exception is Helsenorge in Norway, where only secondary care provides data to the PAEHR. Although all health care personnel in Norway are obliged to provide data to the EHR, it is decided on a regional level whether these data should be electronically accessible in the PAEHR. There are clear similarities among the compared countries when it comes to access points, enrollment, and authentication—all countries use 1 national access point, and each patient of a connected provider automatically receives a PAEHR. Authentication in all countries is made possible through a national electronic ID. Web browsers as well as mobile-adapted web browsers can be used by patients in all studied countries to access the PAEHR. In Norway and Sweden, there is also the possibility of using apps for iOS or Android. There are also some differences regarding how data are provided to the PAEHR. In the case of 1177, Omakanta, and Helsenorge, data are automatically linked to the source EHRs at runtime when a patient logs in to the system. In Digilugu, on the other hand, data are either automatically uploaded from source EHRs to a central server or manually uploaded from the EHR by a health care professional. Hence, in Estonia but not in the other studied countries, it is possible for a health care professional to add content to the PAEHR specifically.

a The questions correspond to questions 2.1 to 2.8 in the data collection template ( Multimedia Appendix 1 , where all response options can also be found).

b EHR: electronic health record.

c PAEHR: patient-accessible EHR.

d FHIR: Fast Healthcare Interoperability Resources.

e ISO: International Organization for Standardization.

f ICD-10: International Classification of Diseases, 10th Revision.

g ATC: The Anatomical Therapeutic Chemical Classification System.

h IHE: Integrating the Healthcare Enterprise.

i XDS: Cross Enterprise Document Sharing.

j HL7: Health Level Seven.

k CDA R2: Clinical Document Architecture Release 2.

l LOINC: Logical Observation Identifiers Names and Codes.

Health Care System Contexts

During the analysis of the collected data, it became clear that it is necessary to provide basic descriptions of the health care system contexts for readers who are unfamiliar with them. Therefore, a summary of the most essential information for understanding the PAEHR context is provided in Textbox 3 . The full data are provided in Multimedia Appendix 3 .

Governance structure

• A parliamentary republic, which is a member of the European Union (EU)
• There are 3 levels of government: state, regions (19), and municipalities (310)
• A parliamentary republic, which is a member of the EU
• There are 3 levels of government: state, counties (15), and municipalities (79)
• A constitutional monarchy, which is not a member of the EU (but a member of the European Economic Area)
• There are 3 levels of government: state, counties (11), and municipalities (356)
• A constitutional monarchy, which is a member of the EU
• There are 3 levels of government: state, regions (21), and municipalities (296)

General health care system financing

• Municipalities are responsible for providing health care that is financed through local tax.
• Every resident in the country is entitled to health care services from the tax-funded system (funds from the 3 levels of government and a small private sector). Municipal authority hospitals provide specialist care, and privately owned hospitals supplement with, for example, day surgery.
• The health care is organized in 3 levels—primary or family care, specialist care, and nursing care
• Payroll tax covers 78% of all expenditures. The Estonian Health Insurance Fund (EHIF) is the sole provider of universal health coverage. The EHIF covers approximately 95% of the population; the rest is covered by the Ministry of Social Affairs. The ministry also has a main responsibility.
• All health care providers are independent entities; family physicians are local entrepreneurs in private companies.
• The health care is organized in 2 levels—primary care (municipalities) and specialist care (the state). Specialist care is divided into 4 health regions and provides specialist care (mainly hospital based) and ambulance service. The state financing is channeled through the health regions.
• The tax-funded system covers health care for all citizens.
• The health care is organized on 3 levels—the state, regions, and municipalities. The regions with their own elected parliaments decide on regional tax, which provides the major financing for health care. National tax funds also constitute large parts of regions’ and municipalities’ health care budgets. Home care is funded through local municipality taxes.
• The tax-funded system covers everyone, including recent immigrants. It is free for inpatient care, and outpatient care has a low cost. Most of the care is operated by regions, but especially primary care and some specialist services are performed by private companies under contracts with the regions. The private share varies a lot among regions, with the largest private providership in the capital region where also a small private part exists independent of the public system that is financed mainly through optional insurance.

Primary care

• These are municipality-arranged services at municipal health centers.
• They include population health monitoring and the promotion of well-being and health as well as prevention, diagnostic services, and treatment.
• Every EHIF-insured individual (and every Estonian resident) is assigned to a personal general practitioner (GP), who is the first level of contact. The insured individual can choose the GP. Approximately 70% of GPs are in solo practices. Practice lists cover the entire population. There are no copayment fees for primary care services in the EHIF package.
• The GP is the main point of contact for health benefits; they are expected to manage most patient pathways and to refer patients to specialist care or long-term nursing care and rehabilitation.
• Provides >50% of ambulatory care visits.
• Services are arranged by the municipalities based on the GP scheme. All citizens have the right to register with a GP of their choice. There are approximately 5000 GPs in total.
• The GPs are the first line of contact; they coordinate care, are gatekeepers for welfare goods, and manage referrals to specialist care.
• Municipalities also offer, for example, emergency care, home care, and rehabilitation services.
• Services are arranged by the regions. Each citizen is connected to a primary care team (often physicians, nurses, psychologists, and physiotherapists), which is the first line of contact. There are approximately 900 primary care centers, between 10% and 50% private depending on the region. Primary care is never provided by single physicians. Care centers are relatively large, with 2 to 10 physicians and a total staff of 20 to 80 people. Organization differs somewhat among regions.
• Primary care handles many issues directly but may also refer to specialist inpatient or outpatient clinics. They also work with health promotion and preventive care (eg, vaccination) and maternity and child care and are the direct clinical contact of the municipalities’ home health care.

Steering of health information and communications technology (ICT) development

• The national solution (Omakanta) is provided by the Social Insurance Institution (Kela). Private and public providers often also provide local portals with no connection to Omakanta.
• Decisions on regional health ICT systems are often made by municipalities or hospital districts. Health Village (portal implemented by all university hospitals) and Omaolo (municipality collaborative) are national initiatives.
• There are 2 main domains: Central databases, services, and applications: the National Health Information System (central database and services governed by the Ministry of Social Affairs) and the database of the EHIF are the major components. Databases and applications of health care stakeholders: electronic medical records (EMRs) and health information systems (HISs) of care facilities, as well as applications provided and maintained through private companies, are the major components. EMRs and HISs need to comply with central systems and legal regulations on, for example, data sharing.
• The Directorate of eHealth was established in 2016 and is responsible for implementing national policy and steering and coordinating eHealth initiatives with stakeholders. They are generally responsible for the work in national eHealth programs, including the national portal Helsenorge. Since 2020, the Norwegian Health Network has had the responsibility to develop and operate the portal. It is the health regions’ decision whether to use the national offered health portal for patient-accessible electronic health records (PAEHRs) in their region.
• Major decisions are made by the 21 autonomous regions; hence, there is not much national coordination. However, collaborations on common procurement of electronic health record (EHR) systems among regions, as well as among secondary care, primary care, and municipalities in some larger regions, have happened recently (Cambio COSMIC is procured by 17 of 21 regions).
• Inera (subsidiary of the Swedish Association of Local Authorities and Regions) supports regions with ICT development and interoperability issues. Inera manages the 1177 patient portal, where Journalen resides.
• The eHealth authority (under the Ministry of Health and Social Affairs) manages electronic prescriptions of medicinal products. They hold the prescription database and link together all pharmacies.

Aims and Motivation

The study presented in this paper had two main aims: (1) to develop a detailed collaborative method suitable for cross-country sociotechnical analyses of eHealth services for patients’ access to their records and (2) to illustrate the data collection method by comparing results regarding contextual factors enabling PAEHR access and use in 4 European countries. Starting from a sociotechnical analysis framework developed by Sittig and Singh [ 17 ] with an intention to produce a detailed data collection template for cross-country sociotechnical comparisons, a collaboration method was developed and implemented. The resulting data collection instrument was then used in a case study where a cross-country analysis was conducted based on collected data from Estonia, Finland, Norway, and Sweden.

This sociotechnical analysis is valuable because, while innovation in PAEHRs is important, there must be a concomitant, ongoing focus on how these tools are integrated into health systems. Digital innovations can fragment care or risk not being used, and examining the social and technical factors pertaining to PAEHRs in different settings is crucial. This will help us understand potential nonadoption; abandonment; and the effects of using such tools in health care on a wide range of stakeholders, including patients and clinicians.

Method Development

Developing a method that involved collaboration among knowledgeable representatives from all countries involved was a necessity for arriving at the desired result—a data collection instrument that enabled a detailed cross-country sociotechnical system comparison. Earlier research based on more limited sociotechnical comparisons has shown that PAEHR systems differ greatly across countries, mostly due to differences in health care policy, underlying care structures, and health ICT initiatives [ 3 ]. To be able to develop a data collection instrument that both covers important aspects of the investigated countries and systems and is relevant in the respective health care system contexts, it was deemed of high importance to involve experts from each country.

The NORDeHEALTH project participants were not experts on every aspect covered, but they had established connections with relevant health authorities and development companies that could provide necessary information through documents or in-person communication. These contacts were necessary for the success of the study, and hence, the continuous contact with these external national agencies and companies was a vital part of the developed collaboration method. On the basis of the experience with the method development, we argue that these external connections need to be established before doing similar research in the future.

The sociotechnical analysis framework by Sittig and Singh [ 17 ] on which this work was based was general and theoretical in the sense that dimensions but not any specific data collection items were presented. The framework was also aimed toward sociotechnical health systems in general and not PAEHR systems or even eHealth services for patients in particular. Hence, a big part of the method was devoted to developing a data collection instrument that would work in this specific context—a cross-country analysis of PAEHR systems in the 4 involved countries. An inevitable consequence of this is that the data collection instrument developed in this study cannot be used to enable sociotechnical analysis of other kinds of health care systems. This being said, a similar method can be used to develop other data collection instruments that include items for the different framework dimensions that are better suited to other kinds of health care systems.

The framework by Sittig and Singh [ 17 ] has indeed been adapted for sociotechnical analysis of health care technology systems but not for cross-country comparisons. As a consequence, an important part of this work was to amend it for the purpose of comparing health care systems across countries. In total, 2 dimensions were added to the framework for this purpose—metadata and health care system context. It is beneficial to include the Metadata dimension in any kind of sociotechnical analysis, but we argue that it is especially important to include this dimension when comparing data from different countries with their own systems, responsible data collectors, and information sources.

The collection of data for the Health care system context dimension occurred after the data collection that was based on the developed data collection instrument. After data collection, this dimension was added as a last step of the iterative development of the method and the data collection instrument in line with the design science research approach followed in this study. As earlier research has shown that PAEHR systems vary across countries due to contextual factors [ 3 ], this dimension is of high importance to consider in a cross-country comparison of the kind carried out in this study. The national setup of primary and specialist care, as well as general digital care infrastructure and steering of health ICT development, can affect the results of all other framework dimensions. Hence, we argue that this dimension is necessary to include in these kinds of studies.

Country Comparison

The sociotechnical cross-country comparisons regarding contextual factors brought to light several differences and similarities among the involved countries. Hardware and software computing infrastructure was the dimension in which most of the similarities could be found. However, it is noteworthy that the use of standards differs considerably across countries, which could potentially give rise to interoperability issues in case, for example, the Nordic countries were to move toward a joint service. It is also of interest to note that physicians in Estonia manually upload some content to the PAEHR, whereas Sweden and Norway only use the automatic link to EHRs at runtime. In Finland, content is also uploaded to the centralized PAEHR but by technical or administrative staff rather than clinicians. This means that health professionals in Estonia can be said to be active users of the PAEHR services. There are some risks associated with being an active user. Aside from the fact that the workload of health care professionals increases, there is also a risk that the personal opinions of health care professionals may affect their willingness to share information and, consequently, the patients’ potential to access their information. Overall, the level of engagement of health care professionals affects whether patients will be able to access some information. In Norway and in some regions in Sweden, patients can only see notes that are signed by a health care professional. Some regions in these countries also have a default delay in publishing notes to patients regardless of whether they are signed. This makes it possible to hide information from patients for some time, for example, by not signing a note.

When it comes to user demographic information, similarities could also be found, but it was difficult to compare data due to differences among the statistical information that the countries provide and how these data are presented. In these kinds of analyses, it is important to be able to compare user statistics on how different groups use the PAEHR services, and this comparison was not possible as the statistics covered different user groups in Sweden and Finland. Estonia and Norway did not (at the time of data collection) collect statistics on individual PAEHR use. The difference in available national statistics on demographic use clearly shows the need to collect user statistics preferably based on comparable user groups and age intervals. This would aid future comparisons as well as evaluations of PAEHR services within countries.

In the Health care system context dimension, several differences could be found regarding health care system setup and financing. While all involved countries have 3 levels of government (state, regions or counties, and municipalities), there are differences in which levels provide most of the funding and are responsible for different levels of care. While the regions are responsible for most health care services in Sweden, the responsibility is more divided in the other countries. Differences are also found regarding steering of health ICT development. While, in most countries, the national PAEHR solution is managed by government-owned companies or insurance institutions, there are differences in other areas of health ICT.

Strengths and Limitations

To our knowledge, this kind of detailed comparison among PAEHR systems in different countries is novel. While Essén et al [ 3 ] compared PAEHR systems in 10 different countries, including Estonia, Finland, Norway, and Sweden, their comparison focused on the effect of differences in regulations on a few aspects of PAEHRs. Our study expands this research by using a complete sociotechnical framework as a base, enabling more detailed comparisons and a more in-depth analysis of similarities and differences across countries. This kind of comparison is important if we want to understand the effects of contextual factors on the realization of PAEHR systems and will also make it possible for systems to learn from each other. This learning takes place not only when analyzing the results from the data collection but also when carrying out the developed collaboration method.

This study used a specific sociotechnical framework as a base, and even though the framework by Sittig and Singh [ 17 ] is tailored toward health systems, it risks constraining the focus to certain dimensions. There are other sociotechnical frameworks that could have been used instead. This being said, the involved researchers did add and change some dimensions to better fit PAEHR systems in particular and the cross-country comparison. Hence, even though a specific framework was used as an important base, it did not completely constrain the focus.

The number of countries involved—and, hence, the number of compared systems—was fairly limited, especially considering the fact that there are several clear similarities among the Nordic countries when it comes to national contexts. Comparisons involving more diverse countries should be conducted when validating the method.

Future Work

This paper presented the results of a sociotechnical cross-country comparison among PAEHR systems in Estonia, Finland, Norway, and Sweden related to contextual factors. The dimensions that were in focus in this paper cover 4/11 of the total dimensions in the data collection instrument. Thus, there are more in-depth comparisons being made based on the data collected in this study. In future work, more dimensions will be included, and the results will be published in subsequent papers.

In future research, it would also be beneficial to increase the number of countries to be compared. The next step in this direction could be to extend the analysis conducted by Essén et al [ 3 ] by using the framework by Sittig and Singh [ 17 ] and the collaboration method presented in this paper. This would not only include a comparison of PAEHR systems from more countries with larger differences in national contexts, but it would also validate the method in a more complex study. These kinds of more expanded comparisons could also use other frameworks as their base. It is also important to note here that digital forms were not developed in this study as the number of systems to be included in the data collection in the 4 involved countries was fairly limited and, hence, the statistical analysis made possible in many digital tools would not add any value. Instead, the data collection was prepared as a Word template to be filled out digitally. In future research including more countries and systems, it would be well worth introducing digital forms and statistical analyses.

The study presented in this paper only focused on PAEHR systems. Even though these systems are in increasing focus in research nowadays, there are many more types of eHealth services for patients as well as other types of health care systems. In future research, we would like to see this collaboration method used for developing data collection instruments related to other kinds of health care systems, such as apps for self-tracking or systems for video visits.

Future research could also usefully focus in greater depth on particular sociotechnical dimensions of PAEHRs in different countries. For example, regarding human-computer interfaces, investigators might explore the highly specific design features of portals, which may augment psychological dispositions to access health information, or (possibly more importantly) offer feedback to providers.

Conclusions

In this work, we aimed to develop and test a method that would enable cross-country sociotechnical analysis of PAEHRs in 4 countries where PAEHR use has reached maturity: Estonia, Finland, Norway, and Sweden. The main artifact produced during this work was a sociotechnical data collection template that was based on a sociotechnical framework for studying eHealth services derived by Sittig and Singh [ 17 ]. The data collection template not only considers the dimensions in the framework by Sittig and Singh [ 17 ] tailored for single-system analysis but also includes parts deemed necessary for cross-country comparisons of several systems. Close collaboration during the process among researchers from all involved countries ensured relevance in all settings.

The data collection template was tested through data collection in the 4 countries and a sociotechnical analysis. Results indicated several important similarities and differences among the 4 countries, clarifying that the process followed and the data collection template enabled an in-depth cross-country sociotechnical analysis of PAEHR systems. This paper presents the first part of the results from the sociotechnical analysis—similarities and differences among contextual factors—and companion articles will delve deeper into the remaining parts.

Acknowledgments

This work was supported by NordForsk through the funding to the Nordic eHealth for Patients: Benchmarking and Developing for the Future (NORDeHEALTH project 100477); by the Strategic Research Council at the Academy of Finland (grants 352501 and 352503); and by the Swedish Research Council for Health, Working Life, and Welfare through the funding to Beyond Implementation of eHealth (project 2020-01229).

Data Availability

The data sets generated and analyzed during this study are not publicly available due to the fact that additional publications based on this data set are being produced, but are available from the corresponding author on reasonable request.

Conflicts of Interest

CB is the Associate Editor of JMIR Mental Health .

Final data collection instrument.

Information sources.

Descriptions of health care system contexts.

• Cijvat CD, Cornet R, Hägglund M. Factors influencing development and implementation of patients' access to electronic health records-a comparative study of Sweden and the Netherlands. Front Public Health. Jun 8, 2021;9:621210. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Hägglund M, DesRoches C, Petersen C, Scandurra I. Patients' access to health records. BMJ. Oct 02, 2019;367:l5725. [ CrossRef ] [ Medline ]
• Essén A, Scandurra I, Gerrits R, Humphrey G, Johansen MA, Kierkegaard P, et al. Patient access to electronic health records: differences across ten countries. Health Policy Technol. Mar 2018;7(1):44-56. [ CrossRef ]
• Walker J, Leveille S, Bell S, Chimowitz H, Dong Z, Elmore JG, et al. OpenNotes after 7 years: patient experiences with ongoing access to their clinicians' outpatient visit notes. J Med Internet Res. May 06, 2019;21(5):e13876. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Moll J, Rexhepi H, Cajander Å, Grünloh C, Huvila I, Hägglund M, et al. Patients' experiences of accessing their electronic health records: national patient survey in Sweden. J Med Internet Res. Nov 01, 2018;20(11):e278. [ FREE Full text ] [ CrossRef ] [ Medline ]
• DesRoches CM, Bell SK, Dong Z, Elmore J, Fernandez L, Fitzgerald P, et al. Patients managing medications and reading their visit notes: a survey of OpenNotes participants. Ann Intern Med. Jul 02, 2019;171(1):69-71. [ CrossRef ] [ Medline ]
• Bell SK, Folcarelli P, Fossa A, Gerard M, Harper M, Leveille S, et al. Tackling ambulatory safety risks through patient engagement: what 10,000 patients and families say about safety-related knowledge, behaviors, and attitudes after reading visit notes. J Patient Saf. Dec 01, 2021;17(8):e791-e799. [ CrossRef ] [ Medline ]
• Bell SK, Mejilla R, Anselmo M, Darer JD, Elmore JG, Leveille S, et al. When doctors share visit notes with patients: a study of patient and doctor perceptions of documentation errors, safety opportunities and the patient-doctor relationship. BMJ Qual Saf. Apr 2017;26(4):262-270. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Zanaboni P, Kummervold PE, Sørensen T, Johansen MA. Patient use and experience with online access to electronic health records in Norway: results from an online survey. J Med Internet Res. Feb 07, 2020;22(2):e16144. [ FREE Full text ] [ CrossRef ] [ Medline ]
• World Health Organization. Digital health platform handbook: building a digital information infrastructure (infostructure) for health. 2020. URL: https://apps.who.int/iris/handle/10665/337449 [accessed 2023-12-22]
• What is an electronic health record (EHR)? HealthIT.gov. URL: https://www.healthit.gov/faq/what-electronic-health-record-ehr [accessed 2023-12-22]
• Metsallik J, Ross P, Draheim D, Piho G. Ten years of the e-health system in Estonia. In: Proceedings of the 3rd International Workshop on (Meta)Modelling for Healthcare Systems. 2018. Presented at: MMHS 2018; June 13-15, 2018; Bergen, Norway.
• Kujala S, Hörhammer I, Väyrynen A, Holmroos M, Nättiaho-Rönnholm M, Hägglund M, et al. Patients' experiences of web-based access to electronic health records in Finland: cross-sectional survey. J Med Internet Res. Jun 06, 2022;24(6):e37438. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Blease C, Salmi L, Rexhepi H, Hägglund M, DesRoches CM. Patients, clinicians and open notes: information blocking as a case of epistemic injustice. J Med Ethics. May 14, 2021;48(10):785-793. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Om pasientjournal. Helsenorge. URL: https://www.helsenorge.no/pasientjournal/om/ [accessed 2023-09-26]
• Baxter G, Sommerville I. Socio-technical systems: From design methods to systems engineering. Interact Comput. Jan 2011;23(1):4-17. [ FREE Full text ] [ CrossRef ]
• Sittig DF, Singh H. A new sociotechnical model for studying health information technology in complex adaptive healthcare systems. Qual Saf Health Care. Oct 2010;19 Suppl 3(Suppl 3):i68-i74. [ FREE Full text ] [ CrossRef ] [ Medline ]
• Hägglund M, Scandurra I. A socio-technical analysis of patient accessible electronic health records. Stud Health Technol Inform. 2017;244:3-7. [ FREE Full text ] [ Medline ]
• Hägglund M. Nordic countries lead new initiative on patient access to EHRs. BMJ Opinion. May 18, 2021. URL: https:/​/blogs.​bmj.com/​bmj/​2021/​05/​18/​maria-hagglund-nordic-countries-lead-new-initiative-on-patient-access-to-ehrs/​ [accessed 2023-12-22]
• Hägglund M, Blease C, Johansen MA, Kujala S, Scandurra I. Exploration of a socio-technical analysis template for patient accessible electronic health records. In: Proceedings of the 2020 Special Topic Conference of the European Federation for Medical Informatics. 2020. Presented at: EFMI STC 2020; November 26-27, 2020; Virtual Event. URL: https://tinyurl.com/349djn9n
• Peffers K, Tuunanen T, Rothenberger MA, Chatterjee S. A design science research methodology for information systems research. J Manage Inf Syst. Dec 08, 2014;24(3):45-77. [ CrossRef ]
• 1177 Vårdguiden. URL: https://www.1177.se/ [accessed 2024-07-02]
• Terviseportaal. URL: https://www.terviseportaal.ee/ [accessed 2024-07-02]
• OmaKanta. Kanta. URL: https://www.kanta.fi/omakanta [accessed 2024-07-02]
• Helsenorge. URL: https://www.helsenorge.no/ [accessed 2024-07-02]

Abbreviations

Edited by A Mavragani; submitted 22.12.23; peer-reviewed by A Mahmud, P Pohl; comments to author 29.02.24; revised version received 05.04.24; accepted 08.04.24; published 26.08.24.

©Jonas Moll, Isabella Scandurra, Annika Bärkås, Charlotte Blease, Maria Hägglund, Iiris Hörhammer, Bridget Kane, Eli Kristiansen, Peeter Ross, Rose-Mharie Åhlfeldt, Gunnar O Klein. Originally published in the Journal of Medical Internet Research (https://www.jmir.org), 26.08.2024.

This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research (ISSN 1438-8871), is properly cited. The complete bibliographic information, a link to the original publication on https://www.jmir.org/, as well as this copyright and license information must be included.

Federated Learning of XAI Models in Healthcare: A Case Study on Parkinson’s Disease

• Open access
• Published: 28 August 2024

Cite this article

You have full access to this open access article

• Pietro Ducange 1 ,
• Francesco Marcelloni 1 ,
• Alessandro Renda 1 &
• Fabrizio Ruffini 1  

Artificial intelligence (AI) systems are increasingly used in healthcare applications, although some challenges have not been completely overcome to make them fully trustworthy and compliant with modern regulations and societal needs. First of all, sensitive health data, essential to train AI systems, are typically stored and managed in several separate medical centers and cannot be shared due to privacy constraints, thus hindering the use of all available information in learning models. Further, transparency and explainability of such systems are becoming increasingly urgent, especially at a time when “opaque” or “black-box” models are commonly used. Recently, technological and algorithmic solutions to these challenges have been investigated: on the one hand, federated learning (FL) has been proposed as a paradigm for collaborative model training among multiple parties without any disclosure of private raw data; on the other hand, research on eXplainable AI (XAI) aims to enhance the explainability of AI systems, either through interpretable by-design approaches or post-hoc explanation techniques. In this paper, we focus on a healthcare case study, namely predicting the progression of Parkinson’s disease, and assume that raw data originate from different medical centers and data collection for centralized training is precluded due to privacy limitations. We aim to investigate how FL of XAI models can allow achieving a good level of accuracy and trustworthiness. Cognitive and biologically inspired approaches are adopted in our analysis: FL of an interpretable by-design fuzzy rule-based system and FL of a neural network explained using a federated version of the SHAP post-hoc explanation technique. We analyze accuracy, interpretability, and explainability of the two approaches, also varying the degree of heterogeneity across several data distribution scenarios. Although the neural network is generally more accurate, the results show that the fuzzy rule-based system achieves competitive performance in the federated setting and presents desirable properties in terms of interpretability and transparency.

Explore related subjects

• Artificial Intelligence
• Medical Ethics

Avoid common mistakes on your manuscript.

Introduction and Motivations

The extensive reliance on artificial intelligence (AI) and machine learning (ML) tools in the healthcare sector poses significant challenges, especially concerning the concept of trust . Any AI system must meet the requirements of robustness, fairness, and transparency throughout its whole life cycle. Furthermore, sensitive health-related data hold an intrinsic value and become a lucrative target for cyber attacks.

The concept of trustworthy AI has recently been considered also by government entities: European Union, for example, is at the forefront for AI regulation as witnessed by the proposal of the “AI ACT” Footnote 1 (2021), which is often referred to as the first law on AI and is conceived for introducing a common regulatory and legal framework for AI. The European Commission had previously promoted the definition of the “Ethics guidelines for trustworthy AI” [ 1 ], which identifies lawfulness, ethics, and robustness as key pillars for trustworthiness and describes the requirements for an AI system to be deemed trustworthy. The ethical aspects are pivotal in the healthcare domain given the sensitive nature of patient data, the disclosure of which poses serious risks. For example, discrimination based on such data can occur in the insurance field: insurance companies could decide to charge different fees depending on the individual health status. Likewise, securing the non-discrimination in financial services is nowadays perceived as an important matter, as witnessed for example by the regulations on the “right to be forgotten” for cancer survivors [ 2 ]. Finally, special attention should be paid to specific domains (e.g., mental health), due to the stigmatized nature of some types of illness.

In the pursuit of trustworthiness, data privacy and transparency emerge as pivotal enablers, especially in the healthcare domain. While data privacy is considered an invaluable right, it somehow clashes with the creation of accurate ML models that, to date, require large amounts of data in their training phase. The common scenario is in fact that many different entities (be they individuals, medical centers, or hospitals) have few or limited amounts of data and are often reluctant to share their assets and sensitive information with other parties. The processes of data mining and knowledge extraction are therefore hampered by the unfeasibility of data collection for centralized processing. The requirement of transparency encompasses the traceability of the learning process, beginning from the data gathering phase, and the ability to comprehend the structure and the functioning of the ML model itself. The latter challenge is the central focus of a branch of AI named Explainable AI (XAI) [ 3 , 4 , 5 , 6 ]. The right for explanation is explicitly mentioned both in the “Ethics guidelines for trustworthy AI" [ 1 ], “ [...] AI systems and their decisions should be explained in a manner adapted to the stakeholder concerned ”, and in the recital 71 of the General Data Protection Regulation (GDPR) [ 2 ] “ [...] the processing should be subject to suitable safeguards, which should include specific information to the data subject and the right to obtain human intervention, to express his or her point of view, to obtain an explanation of the decision reached after such assessment and to challenge the decision.” Indeed, the ability to understand the inner working of an AI system represents a cornerstone of trust and holds particular significance in high-stakes applications in the healthcare domain.

In this work, we embrace the challenge of enhancing trustworthiness of AI systems in medicine investigating technical enablers for the requirements of data privacy and explainability.

Data access limitations, driven by privacy requirements and by the need to prevent ethical risks associated with the disclosure of sensitive data, have prompted the development of new paradigms for training ML models, including federated learning (FL) [ 7 , 8 ]. FL enables multiple parties to collaboratively train an ML model without any disclosure of private raw data. Essentially, a shared global model is learned through proper aggregation of locally computed updates from remote data owners, thus removing the need of centralizing data for training purposes.

The requirement of explainability is typically addressed through two main categories of approaches [ 3 , 4 ]: the exploitation of interpretable by-design models and the adoption of post-hoc explainability techniques. Interpretability and transparency refer to inherent properties of a model and consist in the ability to understand how decisions have been taken and what is the structure of the model itself, respectively. Post-hoc explanation techniques, instead, typically address the goal of explaining why a model provides a decision. It follows that an interpretable model results to be also explainable. The property of interpretability is generally attributed to models such as decision trees (DTs) and rule-based systems (RBSs): in fact, they consist of (or can be traced back to) collection of “IF antecedent THEN consequent ” rules. As a consequence, the inference process turns out to be highly understandable. Understandability can be defined as “ the characteristic of a model to make a human understand its function (that is, how the model works) without any need for explaining its internal structure or the algorithmic means by which the model processes data internally ” [ 3 ]. It is worth emphasizing that the concept of understandability is strongly related to the targeted audience in terms of their a-priori knowledge and cognitive skills. For example, a rule-based inference requires familiarity with logic and, possibly, an adequate training depending on the specific implementation adopted in the antecedent and consequent parts of the rules.

Post-hoc techniques are applied on models which are referred to as opaque or “black boxes,” such as Neural Networks (NNs) and ensemble models, to explain their outcomes. The vast majority of existing post-hoc methods can be roughly ascribed to the following categories, which emulate different nuances of human reasoning: feature importance explanations, rule-based explanations, prototypes explanations, contrastive/counterfactual explanations, and textual or visual explanations [ 3 , 9 ]. Notably, since the field of research is constantly evolving, the list should not be considered exhaustive; in addition, different post-hoc strategies can be specifically tailored based on the different kinds of data (e.g., images or text) and based on the specific models involved [ 9 ]. In the context of XAI, a further distinction is also made between local and global explanations: the former refers to the inference process and focuses on how/why the decision is taken for any single input instance. The latter refers to structural properties of the models (thus pertaining to the concept of transparency) or to aggregated information computed over the entire dataset.

The awareness of the importance of explainability and privacy preservation has greatly increased in recent years. While FL inherently tackles the challenge of preserving data privacy in decentralized ML, it typically lacks integrated solutions for the issue of explainability [ 10 ]. Actually, FL was originally conceived for models optimized through stochastic gradient descent (SGD) (e.g., Deep Neural Networks (DNNs)), and the application of post-hoc techniques is not straightforward in the federated setting.

This work lies at the intersection between FL and XAI and contributes to a research area named Fed-XAI (acronym for fed erated learning of XAI models) [ 11 , 12 , 13 ]. We explore the adoption of Fed-XAI approaches within the healthcare domain for predicting the progression of Parkinson’s disease (PD), formulated as a regression task. We consider a plausible scenario where sensitive raw data originate from different medical centers, making centralized learning unfeasible. In particular, the task is to predict one of the most commonly used indicators for the severity of PD symptoms, namely the Unified PD Rating Scale (UPDRS, firstly introduced in 1987 [ 14 ]), by exploiting real-world voice recordings. The analysis extends a recent work [ 15 ] and encompasses two approaches for Fed-XAI in order to explore the trade-off between accuracy and trustworthiness. The first approach adopts an interpretable by-design model, learned in a federated fashion. The second one employs an opaque model, where both training and post-hoc explanation are compliant with the federated setting. On one hand, we can assess the generalization capability of the models in the regression problem by exploiting a real, publicly available dataset; on the other hand, trustworthiness is meant here as the concurrent attainment of explainability in all its nuances, from both local and global perspectives, and privacy preservation through the adoption of FL.

As for the interpretable by-design model, we employ the Takagi-Sugeno-Kang Fuzzy Rule-Based system (TSK-FRBS) [ 16 ] which is considered as a transparent and interpretable model: its inference method mimics a cognitive process typical of human reasoning in the form of if-then rules . The partitioning of numerical variables into fuzzy sets, which is one of the defining aspects of TSK-FRBS, has proven to enable competitive levels of performance for classification and regression tasks [ 17 ] and has a twofold implication. First, fuzziness in rule-based systems enhances semantic interpretability through linguistic representation of numerical variables. Second, a fuzzy set can be interpreted as a formal representation of an information granule, intended as a generic and conceptually meaningful entity [ 18 , 19 ]: in this context, a fuzzy set allows any number in the real unit interval to represent the membership degree of a feature value to the information granule. As a consequence, the adopted TSK-FRBS fits into the paradigm of granular computing and makes use of information granules in the explainable decision-making pipeline.

As for the opaque model, we employ a well-known biologically inspired model, namely Multi Layer Perceptron Neural Network (MLP-NN). FL is performed exploiting the popular federated averaging (FedAvg) aggregation strategy [ 7 ]. Furthermore, the SHAP explainer [ 20 ], purposely adapted to comply with the federated setting based on a recently proposed approach [ 21 ], is used for explaining the output of the MLP-NN by attributing the contribution (i.e., importance) of each feature to each prediction.

The main contributions of this work can be summarized as follows:

We simulate a scenario in which several medical institutions cooperate in creating a PD progression prediction model pursuing the requirements of explainability and privacy preservation;

To achieve this goal, we implement and exploit two Fed-XAI approaches, based on TSK-FRBS, and MLP-NN plus SHAP, which represent state-of-art techniques for by-design interpretability and post-hoc explainability, respectively, in the federated setting;

We discuss the accuracy of the approaches under several data distribution scenarios, considering the independent and identically distributed (i.i.d.) case and three different non-i.i.d. cases;

For each scenario, we compare the FL scheme with two baselines, namely centralized learning and local learning, to verify the suitability of the federated approach;

We discuss the explainability of MLP-NN and the interpretability of TSK-FRBS, both from a local and a global perspective;

We discuss about the consistency of explanations provided by the two approaches in the federated setting. Here, consistency is achieved when different participants in the FL process obtain the same explanation given the same input information.

The rest of the paper is organized as follows: in Section 2 , we provide a brief overview of recent works that adopt XAI and FL tools in healthcare and more specifically in the context of PD. Furthermore, we describe recent advances in the field of Fed-XAI. Section 3 describes the background related to FL, detailing the approaches for FL of TSK-FRBS and FL of MLP-NN. Furthermore, SHAP is introduced as post-hoc explainability technique, and a recent approach for exploiting SHAP in the federated setting is presented. Section  4 describes the PD progression prediction case study, providing details about the experimental setup: we outline the different data distribution scenarios, the evaluation strategies, and the configuration of the two approaches. In Section 5 , we report and discuss the experimental results. The considerations regarding interpretability and explainability are given in Section 6 . Finally, in Section 7 , we draw some conclusions.

Related Works

The adoption of AI techniques in healthcare has been widely investigated. In this section, we first discuss the most relevant works concerning the adoption of FL and XAI in this application domain. Then, we discuss existing works related to XAI in Parkinson’s disease studies. Finally, we discuss recent algorithmic efforts for combining FL paradigm and XAI approaches.

Federated Learning and XAI in Healthcare Scenarios

The opportunities and the practical utility of FL in the healthcare domain have been recently acknowledged in the specialized literature [ 22 , 23 , 24 ], with applications mainly in the fields of medical imaging [ 25 ] and precision medicine [ 26 ]. FL is presented as a solution to protect sensitive data for privacy concerns and ethical constraints [ 27 ] and also in relation to cyber attacks [ 28 ]. At the same time, the interest in XAI is increasingly widespread, especially in the attempt to “open” the so-called black-boxes [ 29 , 30 ], which have enabled unprecedented performance in the field of deep learning (DL) in medicine. The surveys on XAI for healthcare applications usually delve into the problem of how to present the AI results and their explanations to physicians, medical staff, patients, and caregivers: the explanations should be a tool to understand the outcomes of an AI system, but also a way to allow interaction and enhance stakeholders’ trust in AI (human-centered AI). The XAI goal is usually achieved through the adoption of post-hoc methods for opaque models, often concerning image data analysis (e.g., X-rays and CT scans).

Authors of a recently published survey [ 10 ] provide a review of clinical cases where post-hoc methods and interpretable by-design models are applied to more than 20 different medical case studies, spanning from COVID-19 diagnosis and early detection to diagnostic for breast cancer. Different data types are exploited, depending on the application: images (e.g., EEG, MRI) are often involved, and SHAP is among the most used post-hoc methods. Example case studies include prediction of depressive symptoms from texts with adoption of a post-hoc method for the estimation of word importance [ 31 ] and Alzheimer classification using Random Forest and SHAP [ 32 ].

In the same survey [ 10 ], the practical utility of FL in healthcare applications is discussed, especially considering DL approaches, horizontal data partitioning, and FedAvg optimization strategy. Example case studies include the detection of COVID-19 from decentralized medical data, with Convolutional Neural Networks applied on anterior and posterior chest X-rays [ 33 ]. Authors in [ 34 ] exploit tabular electronic health records (demographics, past medical history, vital signs, lab tests results) from five hospitals to predict mortality in patients diagnosed with COVID-19 within a week of hospital admission. However, it is worth noticing that the applications of FL and XAI are treated separately, emphasizing the substantial lack of works that simultaneously address the requirements of privacy through FL and transparency through XAI in the healthcare domain.

XAI in Parkinson’s Disease Studies

PD is diagnosed in about 10 million people worldwide [ 35 ]: after the Alzheimer, it is one of the most prevalent neurodegenerative diseases. Given its socioeconomic relevance, several AI methods have been proposed for supporting diagnosis and monitoring [ 35 , 36 ]. The most commonly used data types exploited for PD studies include images and speech signals [ 37 , 38 ].

A few works discuss the topic of explainability in the context of PD studies supported by AI techniques: for instance, authors in [ 39 ] apply the LIME [ 40 ] post-hoc method on a DNN used to classify healthy from not-healthy subjects using images from SPECT scanning. Authors in [ 41 ] provide explanations for different ML model outcomes using three post-hoc methods, namely LIME, SHAP, and SHAPASH (a tool for making ML models more understandable and interpretable for general audience), on a multiclass classification task. Since the aspect of data privacy holds high relevance in this context, few recent works elaborate upon the exploitation of the FL paradigm for PD-related applications [ 42 , 43 , 44 ].

In this work, we consider the Parkinson Telemonitoring dataset, which has been analyzed in several recent works for both classification [ 45 , 46 ] and regression [ 47 , 48 , 49 ] tasks. None of the works mentioned above, however, considers the aspects of privacy and explainability simultaneously.

The primary goal of our analysis is to understand the potentialities of the Fed-XAI paradigm in a PD-related application. In the following, we provide a brief overview of the most relevant approaches for Fed-XAI proposed in the literature, relaxing the constraint on the application domain.

Federated Learning of XAI Models

Explainability in FL has been pursued either using post-hoc [ 20 , 50 , 51 , 52 , 53 , 54 , 55 ] or ex-ante [ 13 , 56 , 57 , 58 , 59 ] approaches. A thorough review of such approaches has been provided in several recent works [ 11 , 12 , 60 ]. Here, we describe the most recent advances on the topic.

Bogdanova et al. [ 60 ] have proposed a novel approach (named DC-SHAP) for consistent explainability over both horizontally (different instances, same features) and vertically (different features, same instances) partitioned data for the Data Collaboration (DC) paradigm. This paradigm consists of two stages: first, participants obtain intermediate representations of data through irreversible transformations and transmit them to a central server (unlike FL, which typically shares models rather than data). Then, the server combines such intermediate representations into a single dataset, trains an ML model, and distributes it back to the participants. Unlike other ex-ante [ 58 ] and post-hoc [ 50 , 61 ] explainability approaches tailored for the decentralized setting, DC-SHAP ensures consistency of explanations: In this context, the property of consistency is met if the explanations of the same data instance for a global model are the same for different participants. As underlined by the authors in [ 60 ], model-agnostic post-hoc explainability methods are prone to misalignment of client-side explanations, since they rely on probing the global model with various inputs generated from the local data distribution (typically referred to as background or reference dataset ). In their proposal for horizontally partitioned data, they use a set of auxiliary synthetic data shared among the participants to solve the issue of different background datasets and show how this allows the mitigation of feature attribution discrepancies among the participants. The approach proposed in [ 51 ] is conceived to obtain a consistent global feature attribution score for horizontal FL. A model-specific post-hoc explainability method, namely Integrated Gradients (IGs) [ 62 ], has been adopted for computing feature relevance. The integrated gradients get averaged and thus unified among the clients; however, local explainability is not addressed.

The issue of Fed-XAI for PD has been recently discussed in [ 63 ], with the aim of identifying digital bio-markers for the progress of the disease. Three assumptions constitute the privacy model, considering a scenario with multiple hospitals, each with its own patients: (i) input records and corresponding labels are isolated; (ii) the raw inputs are isolated between patients; (iii) the target labels are isolated between hospitals.

A hierarchical framework is adopted to build the FL model: local FL processes allow to collaboratively train a model among patients in the same hospital, whereas a global FL process aggregates models from each hospital for generating the complete model. An adaptation of SHAP is then adopted as post-hoc method for feature importance explanation. To address the issue of misalignment of client-side explanations, background datasets are generated sampling from a Gaussian distribution: the parameters of such Gaussian distribution (mean and variance) are estimated for each feature in a hierarchical way, by combining the parameters estimated intra- and inter-hospitals. It is shown that the average feature importance computed in the federated fashion is qualitatively similar to but quantitatively different from that obtained in the centralized fashion, where the union of the participants’ training sets can be used as background dataset. Although the proposed method theoretically allows for it, the aspect of local explainability/interpretability is not however discussed.

Authors in [ 21 ] have proposed an approach for obtaining SHAP explanations [ 20 ] in horizontal FL. Specifically, the explanation of an instance prediction made by the federated ML model is obtained by aggregating the explanation of the participants. Such an approach ensures consistency of explanations and is shown to be a faithful approximation of the SHAP explanations obtained in a centralized setting. However, the approach requires that test instances are available to all participants, which may be undesirable or unfeasible in real-world applications where privacy must be guaranteed also at inference time.

In the framework of FL of interpretable-by-design models, TSK-FRBSs [ 13 , 57 , 59 ] and DTs [ 56 , 58 ] have been considered as XAI models to be learnt in a federated fashion. Approaches proposed in [ 57 , 59 ] for federated TSK-FRBS rely on a clustering procedure for the structure identification stage and on a federated adaptation of classical gradient-based learning schemes for adjusting the parameters of the consequent part of the rules. In this work, we consider the approach introduced in [ 13 ], which leads to more interpretable TSK-FRBSs compared to the ones considered in [ 57 ] and [ 59 ]. Additional details on such an approach are reported in Section 3.1 .

As for DTs, the IBM FL framework [ 56 ] supports, among others, a federated adaptation of the ID3 algorithm for horizontally partitioned data. Specifically, an orchestrating server grows a single decision tree by exploiting client contributions based on their local data, in an iterative, round-based, procedure. Similarly, the approach proposed in [ 58 ] allows multiple clients to collaborate in the generation of a global DT by transmitting encrypted statistics, but it refers to the vertical data partitioning scenario. Finally, Polato et al. [ 64 ] have proposed a federated version of the AdaBoost algorithm, posing minimal constraints on the learning settings of the clients, enabling a federation of DTs, and without relying on gradient-based methods.

The categorization of FL approaches is typically based on the data partitioning scheme and the scale of federation. Data partitioning can be broadly categorized into horizontal and vertical settings. In the horizontal setting, training instances from different participants refer to the same set of features, whereas in the vertical setting, the feature set itself is partitioned among participants. The scale of federation refers to the number of participants and is typically classified into cross-silo FL, involving a low number of participants with ample data and computational power, and cross-device FL, where a large number of participants, often represented by smartphones or personal equipment, may feature a relatively small amount of data and computational power.

The PD progress prediction case study discussed in this work pertains to a cross-silo horizontal FL setting. This section reports background information for the two approaches adopted to address this task, which can be ascribed to the Fed-XAI research field: federated TSK-FRBS and federated MLP-NN with post-hoc explainability.

Federated Learning of TSK-FRBS

Let \(\textbf{X}=\left\{ X_{1}, X_{2}, \dots , X_{M}\right\} \) and Y be the set of M input variables and the output variable, respectively. A generic input instance is in the form \(\textbf{x}_{i} = [x_{i,1}, x_{i,2}, \dots , x_{i,M}]^T\) and has an associated output value \(y_i\) . Let \(U_{j}\) be the universe of discourse of variable \(X_j\) and \(P_{j} =\left\{ A_{j,1}, A_{j,2}, \dots , A_{j, T_{j}} \right\} \) be a fuzzy partition over \(U_j\) with \(T_{j}\) fuzzy sets, each labeled with a linguistic term. The term \(A_{j,t}\) indicates the \(t^{th}\) fuzzy set of the fuzzy partition over the \(j^{th}\) input variable \(X_j\) . A TSK-FRBS consists of a collection of fuzzy if-then rules, where the antecedent part of each rule is a conjunction of fuzzy propositions and the consequent part implements a regression model. In case of the commonly used first-order regression model, the generic \(r^{th}\) rule is expressed as follows:

where \(\gamma _{r,j}\) (with \(j=0,\dots ,M\) ) are the coefficients of the linear model that evaluates the output prediction \(y_r\) .

The parameters of the rules are determined through a data-driven approach. The if part (antecedent) is generated either using grid-partitioning or fuzzy clustering over the input space. Once the antecedent is determined, the then part (consequent) estimation consists of local linear models obtained, for instance, through the least squares method.

At inference time, TSK-FRBS exploits the rule base as follows. Given an input instance, first the strength of activation of each rule is computed as

where \(\mu _{j, t_{r,j}}(x_{i,j})\) is the membership degree of \(x_{i,j}\) to the fuzzy set \(A_{j,t_{r,j}}\) . Then, the final output can be evaluated with either weighted average or maximum matching policy. In the former case, the TSK-FRBS output is computed as the average of the outputs of all the activated rules weighted by their strengths of activation. In the latter case, the output corresponds to the output of the rule with the maximum strength of activation.

The maximum matching policy enhances the interpretability of TSK-FRBS, since a single rule explains a predicted output for an input instance. Furthermore, the fuzzy linguistic representation of numerical variables fosters the semantic interpretability of the model itself, whose operation, based on the evaluation of rules, turns out to be highly intuitive.

From an algorithmic perspective, FL of TSK-FRBS, as well as of other families of highly interpretable models, requires ad-hoc strategies. In this work, we rely on the approach for building TSK-FRBSs in a federated fashion recently proposed in [ 13 ]. We consider horizontally partitioned data: every participant produces a local TSK-FRBS and sends it to the server. Subsequently, the server consolidates the received rule bases by juxtaposing the rules received from the participants and by resolving potential conflicts. A conflict occurs when rules from different local TSK-FRBSs have the same antecedent, thus identifying the same specific region of the input space, but they have different consequents. In this case, the federated TSK-FRBS summarizes conflicting rules in a single rule with the same antecedent as the conflicting rules and with the consequent obtained by computing the weighted average of the regression model coefficients in the consequents of the conflicting rules. Such average takes into account the weight associated with each rule, which is estimated on the local training set as the harmonic mean of its support (how many instances activate the rule), and confidence (average quality of the prediction of the rule).

With the aim of ensuring the consistency of the rules among participants and increasing the system interpretability, the input variables are partitioned by using a strong uniform fuzzy partition with triangular fuzzy sets An example of strong uniform fuzzy partition with five triangular fuzzy sets is shown in Fig.  1 . Here, each fuzzy set is associated with a meaningful label that is used to express linguistically the rules.

An example of strong uniform fuzzy partition with five triangular fuzzy sets

It should be noted that building a fuzzy system requires a careful design especially regarding the choice of its hyperparameters (e.g., number, shape, and position of fuzzy sets for each partition), also considering their impact on interpretability [ 65 ]. Uniform fuzzy partitions with triangular fuzzy sets are generally deemed highly interpretable since they satisfy the criteria of coverage, completeness, distinguishability, and complementarity [ 66 ]. However, in practical applications, a meaningful partition should be agreed with the human users who are expected to interact with the AI system and to interpret the given rules. An interesting future development of this work would consist in examining the choice of the uniform partitioning together with domain experts (e.g., physicians).

Federated Learning of MLP-NN

Models optimized through Stochastic Gradient Descent (SGD), such as NNs, can be learned in a federated fashion by exploiting an aggregation strategy based on, or derived from, the popular federated averaging (FedAvg) procedure. FedAvg is an iterative, round-based procedure, in which each round encompasses the following steps: the server sends the global model to the participants; each participant locally updates the model through SGD on its local training set and sends the updated model to the server; the server obtains an updated global model by computing the weighted average of the locally updated models, where weights are based on the local training set cardinality. Several extensions of FedAvg have been proposed in the literature, mostly aimed at addressing FL in heterogeneous settings [ 67 , 68 ]. In this paper, we focus on classical FedAvg and deal in more detail with the issue of explainability of MLP-NN learned in a federated fashion. To this end, in the following, we first describe a popular post-hoc method for explainability, namely SHAP (SHapley Additive exPlanations) method [ 20 ], and then an approach for the adoption of SHAP in the federated setting.

Post-hoc Explainability: The SHAP Method

One of the most popular post-hoc strategies used to explain a model prediction is to assess the importance of each feature in producing the output. In general, given a model f , an input instance \(\textbf{x}_{i}\) , and a predicted output \(\hat{y_i} =f(\textbf{x}_{i})\) , the explainer assigns to each input component \(x_{i,j}\) a value that reflects how much that particular feature is important for the prediction. These values are interpreted in terms of sign and magnitude: if the sign is positive (negative), then that feature contributes positively (negatively) to the prediction output; as per the magnitude, the larger it is, the higher the impact of the corresponding feature on the output.

In this work, we adopt the SHAP method [ 20 ] which is one of the most widely used approaches to assess the feature importance for both regression and classification tasks. SHAP provides local explainability, that is, it explains individual predictions. Global explainability insights can be obtained by aggregating the individual explanations over a set of data.

SHAP computes the importance of the individual features using the optimal Shapley values introduced by L. Shapley in 1953 [ 69 ] in game theory. In SHAP, the connection between game theory and explainability is that a prediction for an individual instance \(\textbf{x}_{i}\) can be explained by conceiving the features \(X_i\) as the “players” of a “game” where the prediction \(\hat{y_i}\) is the game “payout.” Intuitively, the different M players of the game (features) receive different rewards, called Shapley values \(\phi _j\) , depending on their contribution to the total prediction, i.e., \(\hat{y_i} = \phi _0 + \sum _{j=1}^M \phi _j\) where \(\phi _0\) is a reference value (baseline) computed as the average of output values. In this game-explanation analogy, the player who contributes with the larger \(\phi _j\) to the total prediction is the most important feature in the explanation.

Since the computation of the Shapley values involves testing all the possible combinations of the features ( coalitions of the players in the game theory) by perturbating the instance \(\textbf{x}_i\) , the time increases exponentially with the number of features [ 70 ]. Thus, various approaches have been proposed to estimate them efficiently, including SHAP. There are several kinds of SHAP methods, corresponding to different ways of approximating the Shapely values. In this work, we consider the widely adopted KernelExplainer variant of SHAP (KernelSHAP), as it is model-agnostic [ 21 ]. Indeed, other methods, such as TreeExplainer, result to be more efficient but are model-specific. Algorithm 1 describes the KernelSHAP procedure.

KernelSHAP algorithm, from [ 70 ].

Notably, KernelSHAP requires a background dataset that serves as a reference: whenever a feature is excluded from a coalition, its value is replaced using an instance randomly sampled from such dataset. The choice of a representative background dataset is crucial for obtaining accurate estimates of the Shapley values. For this reason, the training set is typically adopted for this purpose. However, it is not the unique possible choice: a different, generally smaller, dataset can be used at the condition of being representative of the data distribution of the training set. In the literature, representative objects such as medoids or centroids of clusters generated by applying a clustering algorithm on the training set have allowed faster estimations of the Shapley values. Additional details on SHAP and Shapley values can be found in [ 70 ].

Federated SHAP

Let H be the number of clients involved in the federation, \(\textbf{x}_{i} { aninputinstance},\,{ and}f(\textbf{x}_{i}){ thepredictiontobeexplained}.{ Inthiswork},\,{ themodel} f(\cdot ) \) is an MLP-NN learnt in a federated fashion. Following the setup proposed in [ 21 ], the goal is to achieve a federated explanation for \(f(\textbf{x}_{i}){ consideringtheinputinstancesimultaneouslyavailabletoalltheclients}.{ Itisworthunderliningthatthismaybeundesirableorunfeasibleiftheinstance}\textbf{x}_{i} { issubjecttotheprivacyconstraint}.{ Notably},\,{ insomeparticularscenarios},\,{ forexample},\,{ ifthepatientsrequiremultiplemedicalconsultations},\,{ theinstance}\textbf{x}_{i} \) could, under specific agreements, be shared to the other clients. Another scenario is represented by the presence of a benchmark dataset available to research entities, with the objective of comparing the goodness of the explainability produced by different methods.

As discussed in Section 3.2.1 , the application of the SHAP method requires a background dataset. Typically, this is the dataset used to train the prediction model. However, in the federated setting, local training sets belong to different entities and cannot be shared due to privacy issues. Therefore, the server has no reference data to be used as background. Federated SHAP proposed in [ 21 ] represents a possible strategy to overcome this issue and to achieve federated explanations by exploiting the additive property of the Shapley values.

In the federated SHAP procedure, first of all, each client estimates the Shapley values using the local dataset as background dataset; then the values are transmitted to the server that evaluates their average. In this way, the data privacy is preserved, since the raw data are never shared. Furthermore, it is shown that the average can be considered a good approximation of the Shapley values calculated if the union of the local training sets was available to the server.

Schematically, the overall procedure for FL with post-hoc explainability technique entails the following steps: (i) each participant \(h ({ with}h=1, \dots , H){ contributestothecreationofanFLmodel}.{ Attheendofthefederatedprocedure},\,{ thefederatedmodelismadeavailabletoeachparticipant};\,({ ii}){ givenanunseeninstance}\textbf{x}_{i},\,{ eachparticipant}h{ computestheShapelyvalues}\phi ^{(i,h)}_j{ with}j = 1, \dots , M,\,{ evaluatingKernelSHAPlocally},\,{ byexploitingtheprivatetrainingsetasbackgrounddataset};\,({ iii}){ theShapelyvalues}\phi ^{(i,h)}_j{ estimatedatparticipantlevelaretransmittedtotheserver},\,{ whichperformssimpleaveragingtoobtainthefederatedestimationoftheShapleyvalues}\phi ^{(i)}_j{ forexplainingtheprediction}f(\textbf{x}_i)\) .

Case Study: Federated Learning for Parkinson’s Disease Progress Prediction

In this section, we describe the case study for the evaluation of the Fed-XAI approaches. First, details about the PD telemonitoring dataset are provided. Then, we describe the experimental setup in terms of data partitioning scenarios and evaluation strategies. Finally, we give the configurations of the ML models adopted in the two Fed-XAI approaches based on MLP-NN and TSK-FRBS, respectively.

The Parkinson Telemonitoring Dataset

The Parkinson Telemonitoring dataset is a well-known regression dataset available within the UCI Machine Learning Repository [ 71 ]. The dataset is composed of 5875 instances of biomedical voice measurements from 42 patients with early-stage PD. Data are acquired remotely during a 6-month trial. Each instance corresponds to one voice recording, characterized by 22 features as reported in Table 1 . The regression task consists in predicting the total Unified PD Rating Scale score ( total_UPDRS ) associated with a given voice recording. Differently from motor_UPDRS , which is related only to the motor symptoms, total_UPDRS is related to the overall set of symptoms.

Federated Learning Scenarios

In this paper, extending the preliminary analysis performed in [ 15 ], we consider the challenging setting in which the raw dataset is not available on a single node for centralized processing, as in traditional ML, but it is instead scattered over multiple physical locations, e.g., hospitals or healthcare institutions. Specifically, we simulate several scenarios featuring 10 hospitals (cross-silo FL setting), in order to evaluate the performance of two Fed-XAI approaches under different horizontal data partitioning schemes that could be encountered in real-world situations.

In the following, we formally define the four scenarios considered in our experimental analysis. Let \(P_h(\textbf{x},y){ bethelocaldistributionofinputdata}\textbf{x} { andassociatedtargetvalues}y\) ( total_UPDRS ) for the hospital \(h,\,{ and}P(\textbf{x},y)\) the overall data distribution.

Scenario IID It is a simple independent and identically distributed ( i.i.d. ) setting; formally,

The training data of the ten hospitals follow the same distribution, with about 500 instances each.

Scenario NIID-Q (acronym for n on- i . i . d . q uantity skew). It is a non-i.i.d. setting with quantity skew [ 72 ]: different hospitals can hold different amounts of training data, which follow the same overall distribution.

Scenario NIID-F (acronym for n on- i . i . d . f eature skew). It is a non-i.i.d. setting with feature distribution skew [ 72 ] based on the \(age\) feature; formally,

Each hospital contains training data from only a specific range of ages (e.g., 56 to 57, 58 to 59, \(\dots \) , more than 75 years old). In this scenario, we aim to have training sets with as similar amount of data as possible.

Scenario NIID-FQ (acronym for n on- i . i . d . f eature and q uantity skew). It is a non-i.i.d. setting with both quantity skew and feature distribution skew based on the \(age\) feature. Each hospital contains training data from only a specific range of ages; furthermore, different hospitals can hold different amounts of data.

The four scenarios concern different partitioning schemes for training data. As for the testing data, we consider the case of an external publicly available test set, valid for all the scenarios. The test set follows the overall data distribution (i.e., representative of all age groups) and has 588 instances. The distribution of the training data in the four scenarios is summarized in Table 2 and in Fig.  2 .

As for the NIID-F and NIID-FQ scenarios, it is worth underlining that other features besides age may be affected by bias or skewness. However, this contingency still meets the definition of feature distribution skew. Therefore, the four scenarios enable a thorough and extensive evaluation of the performance of the two Fed-XAI approaches based, respectively, on MLP-NN and TSK-FRBS.

Barplot of the training data partitioning scheme over ten hospitals in the four scenarios. Data for different ages are represented in different colors

Evaluation Settings

Typically, the performance evaluation of a model in the federated setting is performed not only in absolute terms, but also comparatively against two baseline settings [ 15 , 73 , 74 ]: local learning and centralized learning. Figure  3 provides a schematic overview of the three learning settings.

Schematized representation of the three learning settings: a federated learning (FL), b local learning (LL), c centralized learning (CL)

Federated learning (FL), local learning (LL), and centralized learning (CL) can be summarized for the dataset under consideration as follows:

FL: the hospitals collaborate in obtaining a single federated model without sharing their raw data. The privacy of sensitive data is preserved.

LL: each hospital locally learns a model from its private training data. As a consequence, the privacy of sensitive data is preserved, as in the FL case, but there is no collaboration among different hospitals.

CL: training data from all hospitals are collected in a single central repository in the server and exploited to learn a global model. CL implies indeed maximum collaboration among hospitals, but violates privacy, as private sensitive data are moved from their owner to the server.

A model learned in the FL setting is expected to be more accurate than the ones learned in the LL setting. On the other hand, a model learned in the CL setting can outperform the other models (both LL and FL), in terms of accuracy, because it can rely on the union of the training datasets. The CL approach, however, is not viable in real applications where privacy protection is a mandatory constraint.

Regression Problem and Fed-XAI Models

The PD progress prediction is formulated as a regression problem where the target variable is the Total_UPDRS score. In our experiments, we replicated the preprocessing steps adopted in [ 15 ], namely, (i) a robust scaling using 0.025 and 0.975 quantiles is applied to the input features to remove outliers and clip the distribution in the range [0,1], and (ii) the output variable is normalized in the range [0,1].

Unlike in [ 15 ], the univariate feature selection procedure is carried out independently for the three learning settings, for a fair comparison of the entire regression pipelines. We select the \(G=4\) best features in terms of Mutual Information (MI) [ 75 ] with the target variable. The estimate of MI and the subsequent feature selection is done individually by each participant in the LL setting, based on the local training sets, and globally in the CL setting, based on the union of the training sets. As for the FL setting, the federated feature selection procedure is schematized in Fig.  4 , considering the example of the IID scenario. Each participant computes the MI score for all the features and transmits such information to the server. The server computes the average MI score for each feature and communicates the \(G\) best features to each participant. Thereafter, the FL process starts considering only the selected subset of features. In the example of Fig.  4 concerning the IID setting, the federated feature selection procedure selects the following features: age , test_time , DFA , and HNR . Note that the feature importance scores of each participant may change depending on the data distribution scenario: thus, the selected features may vary and may generally differ from the CL setting.

A schematic overview of the federated feature selection procedure. The example concerns the IID scenario

The choice of the \(G\) value is guided by the following considerations: a reduced number of features generally improves the explainability task, both for post-hoc and interpretable by-design approaches. In addition, TSK-FRBSs struggle to handle high dimensional datasets [ 76 ]: the set of candidate rules grows exponentially with the number of features, thus jeopardizing the accuracy and the interpretability of the system. We have verified that \(G=4\) ensures a good generalization capability for both models and an increase in the number of features does not lead to a significant improvement in performance.

In our experiments, for each data distribution scenario, we trained a TSK-FRBS and an MLP-NN according to FL, LL, and CL settings. The experimental analysis is approached from a twofold perspective: model accuracy and model explainability . We assess the accuracy of the predictions obtained by the regression models as in [ 15 ] by using two popular metrics, namely Root Mean Squared Error (RMSE) and Pearson correlation coefficient ( \(r\) ). They are defined as follows:

where \(N{ isthenumberofsamplesconsideredfortheevaluation},\,{ and}y_i{ and}\hat{y}_i{ arethegroundtruthvalueandthepredictedvalueassociatedwiththe}i\) -th instance, respectively. Finally, \(\bar{y} { isthemeanofgroundtruthvalues},\,{ and}\hat{\bar{y}} { isthemeanofthepredictedvalues}.{ Obviously},\,{ thegoalistominimizeRMSEandmaximize}r\) .

It is worth underlining that the evaluation of an FL system typically covers other aspects besides accuracy such as computation and communication efficiency. These aspects, however, represent often crucial requirements or potential bottlenecks in cross-device FL, with many devices featuring limited computational resources [ 77 ]. In a cross-silo FL scenario, as the one considered in this work, such aspects are generally deemed less critical.

Interpretable By-design Fed-XAI: TSK-FRBS Configuration

As in [ 15 ], we employ a first order TSK-FRBS model (described in Section 3.1 ). We adopt a strong uniform fuzzy partition on the features with five triangular fuzzy sets, as shown in Fig.  1 . The choice of five fuzzy set is driven by the indication of the specialized literature and by the pursuit of a reasonable trade-off between model complexity and generalization capability. The number of linguistic terms associated with a linguistic variable should be below the limit of \(7\pm 2\) [ 78 ]. Indeed, it has been shown that this represents a threshold for information processing capability, and thus exceeding it undermines the interpretability of the system [ 79 ]. With the aim of describing linguistically a given rule, the five fuzzy sets can be labeled with the following linguistic terms: VeryLow , Low , Medium , High , and VeryHigh . Furthermore, it should be noted that different features may be partitioned using a different number of fuzzy sets, e.g., by exploiting domain knowledge to enhance understandability. Although this represents an interesting future development, we did not conduct extensive hyperparameters optimization. Rather, we verified in the CL setting that beyond 5 fuzzy sets, there is a substantial increase in model complexity, without significant improvement in terms of performance metrics. The choice of 5 fuzzy sets ensures a high linguistic interpretability and represents a reasonable trade-off between model complexity and generalization capability.

TSK-FRBS: empirical cumulative distribution function (ECDF) of the differences of RMSE scores between FL and LL for the four data partitioning schemes

Post-hoc Explainable Fed-XAI: MLP-NN Configuration

The MLP-NN consists of two hidden layers with 128 neurons, each with the ReLu activation function. The Mean Squared Error (MSE) is adopted as loss function and Adam as optimizer. The minibatch size is set to 64. The overall number of epochs is set to 100 in the CL and LL settings. In the FL setting, we set the number of local epochs and the number of rounds as 5 and 20, respectively. We have not performed a thorough optimization of hyperparameters for each learning setting and each data distribution scenario individually; however, we have empirically observed that a further increase in the capacity of the models in terms of number of layers, number of neurons, and training epochs does not lead to a significant increase in the generalization capability of the MLP-NN.

Analysis of the Experimental Results

Table 3 presents the RMSEs and \(r\) coefficients obtained by the TSK-FRBS and MLP-NN models for all the learning settings and the data distribution scenarios. As regards the LL setting, we report the average values \(\pm \) standard deviation obtained by the models learned locally in each participant. In the table, we have highlighted in bold the best results for each row, considering the comparison between TSK-FRBS and MLP-NN. Notably, in certain cases, the result is obtained from a distribution of values and is expressed in terms of mean and standard deviation: this occurs in the LL setting when ten local models are evaluated on the test sets and, regardless of the learning setting, when performance metrics are evaluated on ten training sets from as many hospitals. In such cases, we highlighted in bold the best result only if there exists a statistical difference in metrics values between TSK-FRBS and MLP-NN. The statistical significance has been assessed through a pairwise Wilcoxon signed-rank test [ 80 ] with confidence level \(\alpha =0.05.{ Ingeneral},\,{ federatedmodelsoutperformthelocalcounterparts},\,{ bothintermsofRMSEand}r\) . The benefit of FL over the LL setting is particularly evident for the TSK-FRBS and especially in the non-i.i.d. settings.

Federated TSK-FRBS and federated MLP-NN achieve comparable performances. The most noticeable difference occurs in the NIID-F setting, in which the two metrics provide diverse insights: on one hand, RMSE indicates that the deviation of predictions from true values is lower for the MLP-NN (10.268) compared to the TSK-FRBS (16.848); on the other hand, predictions and true values are more correlated for TSK-FRBS ( \(r=0.461\) ) than for MLP-NN ( \(r=0.205\) ).

The non-i.i.d. setting with quantity skew (NIID-Q) does not harm particularly the performance of the models: RMSE and \(r\) values are comparable to those of the IID setting, for both TSK-FRBS and MLP-NN. In the case of TSK-FRBS, the simple average of the performances measured on the training sets shows poor performance for the LL setting, where some models likely suffer from low data availability. The aggregation strategy based on the rule weight (defined as a combination of confidence and support) ensures that this unfavorable situation is mitigated in the FL setting.

Scenarios with feature distribution skew (NIID-F and NIID-FQ) turn out to be the most challenging for both models. The generalization capability of models built in the LL setting is rather poor, due to exposure to data from a limited age range during training: both TSK-FRBS and MLP-NN perfectly model the training data (resulting in low RMSE values and an \(r\) coefficient around 0.9 on training sets) but fail in properly predicting the total_UPDRS score of the test instances (resulting in high RMSE and low \(r).{ DiscrepancyinRMSEand}r\) values between training and test sets is noticeable in the LL setting, whereas it is limited or negligible in the FL setting.

It is worth highlighting that the performances in the FL setting are always worse than those obtained in the CL setting, both for TSK-FRBS and MLP-NN. In general, the centralized MLP-NN is able to achieve the best performance with a slight improvement in terms of RMSE and \(r\) over the centralized TSK-FRBS. The superior performance of the centralized model can be attributed to the utilization of all data for conventional training. However, it is deemed unfeasible when privacy preservation represents a critical requirement.

The results in Table 3 provide an aggregate view of the LL setting. A better understanding of the outcomes can be gained by analyzing the specific performance obtained in each hospital: such detailed results are illustrated in Figs.  5 and 6 through the empirical cumulative distribution function (ECDF) for the RMSE metric.

MLP-NN: empirical cumulative distribution function (ECDF) of the differences of RMSE scores between FL and LL for the four data partitioning schemes

Number of rules of the TSK-FRBS for each learning setting and each data partitioning scheme. Error bar represents the standard deviation

For both models, the ECDF is reported for the values of the difference, for each hospital, of the RMSE achieved in the FL setting and the 10 locally computed values of RMSE obtained in the LL setting: each plot, therefore, is made up of 10 points. The plot can be interpreted as follows: if a point lies in the negative half-plane (negative RMSE difference), then the RMSE value of the FL model is lower (and therefore the FL model is better) than the RMSE value of the LL model. The fine-grained analysis shown in Figs.  5 and 6 confirms that the FL setting generally outperforms the LL setting.

Finally, we report on the overall complexity of the models, which will be further discussed in Section 6 . In the case of the MLP-NN, the network architecture is fixed: the complexity, intended as the number of parameters, does not change with the learning setting. In the case of TSK-FRBS, the complexity can be assessed in terms of the number of rules in the rule base. Figure  7 shows the complexity for each data distribution scenario and each learning setting.

The number of rules of the federated TSK-FRBS never exceeds, by construction, that of the centralized TSK-FRBS, which in any case is limited (433 rules). As expected, the federated TSK-FRBS is more complex than locally learned ones. In the IID and NIID-Q scenarios (in which each hospital has data representative of the entire distribution), the 10 local models have approximately half the number of rules of the federated ones, indicating that common antecedents are often found in the rule aggregation phase. The complexity of locally learned TSK-FRBSs in the presence of feature distribution skew (NIID-F and NIID-FQ) is significantly lower than the one in IID setting, and the gain in accuracy provided by the federated models comes at a cost in terms of number of rules, which is approximately five times higher.

Explainability Analysis

The extensive adoption of AI systems in the healthcare field depends not only on the achievement of adequate levels of accuracy, but also on how much they are perceived to be trustable. In particular, the ability to explain how the outcomes have been produced by the models is more and more required and represents the main driver of XAI. This section discusses the aspect of transparency of AI systems, focusing on how it is defined for the two Fed-XAI approaches analyzed in this paper. First, we analyze the explainability of the MLP-NN model, in which SHAP is used as post-hoc method. Then, we discuss the interpretability by-design of the TSK-FRBS. We consider models built in a federated fashion according to the IID data partitioning scheme: the discussion of the outcomes is limited to such case, but the pipeline for explainability analysis can be easily replicated for any data distribution scenario. Furthermore, we recall that all the input variables and the output variable are normalized in the unit interval \([0,1]\) : the considerations in this section refer to the predicted values before inverse transformation.

Post-hoc Explainability of MLP-NN

In this section, we discuss the explainability of the MLP-NN after the application of the agnostic post-hoc method SHAP. We recall that, given an input instance, the Shapley value associated with each feature represents the contribution given by such feature to the predicted value. In this sense, for each prediction, SHAP explains why the model produces a particular output.

We adopt the Federated SHAP approach proposed in [ 21 ] and introduced in Section 3.2.2 . KernelSHAP is employed in each hospital to estimate the Shapley values considering the full local training set as background dataset.

It is worth underlining a first crucial aspect concerning the explainability of MLP-NN: unlike interpretable-by-design Fed-XAI approaches, the post-hoc method affects the overall efficiency of the systems, both from a computation and a communication point of view. As for the former aspect, the estimation of the Shapley values with KernelShap is time consuming and the runtime increases with the number of features and the size of the adopted background dataset [ 70 ]. As for the latter aspect, Federated SHAP introduces a communication overhead, as Shapley values need to be transmitted by the participants for central aggregation. Conversely, the interpretable-by-design TSK-FRBS has no computation and communication overhead for generating the explanations.

MLP-NN: Global Insights

Globally, an MLP-NN is generally considered “opaque,” due to the presence of several layers of non-linear information processing. In our case, the structure consists of two hidden layers with 128 neurons, resulting in 17,281 trainable parameters. The high number of parameters and the relations among these parameters make very hard to provide a global explanation of the model. Thus, indirect methods based, for instance, on feature importance are typically used to provide global explainability information [ 3 ].

As Shapley values represent additive feature importance scores for each particular prediction, the overall feature importance can be assessed by computing the average of the absolute Shapely values across the data [ 70 ]. Of course, the larger the average absolute value of the contribution given by a feature, the greater the importance of that feature. The assessment of the feature importance of a model is typically independent of the test data. In the case of the MLP-NN, it can be estimated as follows. First, each hospital \(h{ evaluatestheimportance}I{ ofthefeature}i\) on its training data:

where \(\phi _j^{(i,h)} { istheShapelyvalueforafeature}j{ andatraininginstance}i{ athospital}h,\,{ and}N_h{ isthesizeofthetrainingsetathospital}h\) . Then, each client can transmit locally computed features importance to the server, and the overall features importance for the federated model can be computed by the server as follows:

where \(H{ isthenumberofhospitalsand}N = \sum _{h=1}^{H}N_h\) .

Figure  8 shows the global feature importance scores for the MLP-NN, as per ( 10 ): in the IID setting, the most relevant feature is age , while test_time seems to be less relevant than vocal features, namely DFA and HNR , which in turn are of similar relevance.

Feature importance scores for the MLP-NN

MLP-NN: Local Insights

Figure  9 reports the SHAP values for two instances of the test set; they correspond to two cases where both models (MLP-NN and TSK-FRBS) obtain high and low errors, respectively. The absolute error (AE) made by the MLP-NN is around 0.55 for instance #2496 and around 0 for instance #2323.

MLP-NN local explainability: Shapley values for two instances in the test set. The absolute error for each instance is reported along with the baseline value

The Shapley values reveal, a-posteriori, the relevance of the corresponding features in the prediction performed by the model: as expected, they are different when considering different instances. In the former case (Fig.  9 a), age has little negative influence, while the other features have a large and positive impact on the output value. In the latter case (Fig.  9 b), the most influential feature is test_time , which has a negative impact on the output. Notably, the SHAP values for individual features are evaluated, for both instances, with respect to the same baseline value \(\phi _0 = 0.46\) .

Interpretability By-design of TSK-FRBS

TSK-FRBSs are often considered as “ light gray box” models [ 81 ]: their operation is highly interpretable, since they consist of a collection of linguistic, fuzzy, if-then rules. However, in the first-order TSK-FRBSs used in this paper, the adoption of a linear model in the consequent part, which certainly improves the accuracy with respect to the zero-order TSK-FRBS, makes the interpretation of a single rule less intuitive than the zero-order counterpart.

A substantial difference with respect to the MLP-NN model analyzed in Section 6.1 is that interpretability information is given without additional overhead in terms of computation and communication (as it is the case for the calculation of Shapley values on the MLP-NN model).

The following analysis aims to characterize both global and local interpretability of TSK-FRBS learnt in a federated fashion.

TSK-FRBS: Global Insights

The global interpretability of TSK-FRBSs can be quantitatively assessed by measuring the complexity of the system in terms of the number of rules and/or parameters. Less complex models, i.e., those with fewer rules, can be generally considered more interpretable [ 66 ]. Figure  7 reports the number of rules for each learning setting and each data partitioning scenario. As underlined in Section 5 for the IID case, the number of rules in the FL setting is rather limited (i.e., 397) and just double that in the LL case (despite the presence of 10 participants).

The model is therefore comprehensively described by the rule base, which can be represented in the intelligible form reported in the following.

The number of parameters of a rule in a TSK-FRBS can be estimated as follows, considering the presence of four input variables: it is given by the sum of four parameters for the antecedent part (one for each input variable, to identify a fuzzy set of the a-priori partitioning) and five parameters for the linear model of the consequent part. Ultimately, the considered TSK-FRBS has 3573 parameters overall.

In the case of TSK-FRBS, a measure of feature importance can be obtained by averaging the absolute values of the coefficients of the linear models in the rule base. Formally, the importance \(I{ ofthefeature}j\) is evaluated as follows:

where \(\gamma _{r,j} { isthecoefficientforthevariable}j{ intherule}r\) .

Feature importance scores for the TSK-FRBS

Figure  10 suggests that feature importance computed for the TSK-FRBS model is consistent with the one computed for the MLP-NN by the SHAP method in the IID setting. Age and test_time are identified as the most and the least relevant features, respectively. Furthermore, the importance of HNR and DFA is similar, consistently with what is observed for the MLP-NN.

A direct comparison of the importance values between Fig.  8 (MLP-NN) and 10 (TSK-FRBS) is not meaningful. Indeed, in the case of MLP-NN, Fig.  8 shows the average absolute contribution for each feature with respect to the baseline value. In case of TSK-FRBS, Fig.  10 represents the average absolute value of the coefficients of the linear models used in the case of TSK-FRBS.

It is worth underlining that model-agnostic nature of SHAP can be exploited to compute post-hoc explanations also on the TSK-FRBS model. Thus, we can directly compare the feature importance obtained by averaging the absolute values of the coefficients of the linear models in the rule base and reported in Fig.  10 , with that obtained by averaging the absolute Shapely values across the data for TSK-FRBS ( 9 ) and ( 10 ). The latter approach results in the scores reported in Fig.  11 .

Feature importance scores for the TSK-FRBS evaluated in terms of Shapley values

It is interesting to note that age and test_time are still identified as the most and least important features, indicating a summary agreement among the results obtained with different important attribution methods in the IID scenario. There is a discrepancy, on the other hand, between the relative values of DFA and HNR , possibly because the two approaches estimate importance values with different criteria. Furthermore, it is worth noting that feature importance scores computed for TSK-FRBS using SHAP are consistent with those computed for MLP-NN and shown in Fig.  8 , also in terms of range of values.

The identification of age as the most important feature regardless of the model and the attribution method adopted is not surprising. First, in the IID scenario, each hospital can have data for all age ranges and therefore the feature entails a high variability. Second, such evidence is reflected in the specialized literature, which indicates that age is the best predictor of the progression of Parkinson’s disease and the most important risk factor for the development of the disease [ 82 ].

TSK-FRBS: Local Insights

The interpretability of the TSK-FRBS derives from its structure and the type of inference strategy we use. Indeed, for any given input instance, the predicted output depends on a single rule: the antecedent part isolates a region (a hypercube) of the input space, where its consequent part defines a local linear model. The coefficients of this model indicate how the input features contribute to form the TSK-FRBS output in that region: a positive (negative) coefficient for a given feature indicates that the output increases (decreases) with that feature. Notably, all instances belonging to that region will refer to the same linear model.

Given an instance \(\textbf{x}_{i} { andarule}r,\,{ whichistheonewiththemaximumstrengthofactivationfor}\textbf{x}_{i} \) , the actual contribution of each feature to the prediction \(\hat{y_{i}} { isobtainedas}\gamma _{r,j}\cdot x_{i,j} \) . In other words, contributions are obtained as the element-wise product between the coefficients of the linear model and the feature values of an instance. Figure  12 reports both the feature contributions and the coefficients of the linear model of the TSK-FRBS for the same two instances of the test set analyzed in the case of the MLP-NN (instances #2496 and #2323).

TSK-FRBS local explainability: coefficients of the linear model and actual feature contributions for two instances in the test set. The absolute error for each instance is reported above the relevant plot along with the ID of the rule considered for the prediction and the term \(\gamma _0\) of the linear model

The two instances activate different rules: as a consequence, the contributions are significantly different. Furthermore, in general, the contributions are reduced compared to the coefficients, since each feature is normalized in the unit interval.

Shapley values calculated for the MLP-NN model on the instances in the test set that activate rule \(R_{233} \) of the TSK-FRBS. The absolute error (AE) for each instance and for each model is reported above the relevant plot

Local Explanations: Comparison Between TSK-FRBS and MLP-NN

A relevant outcome can be drawn by comparing the barplots of Figs.  9 and 12 : although the absolute errors are similar (we have verified that the predicted values are similar as well), the two models “reason” differently and assign different—sometimes diametrically opposed—contributions to the features, also because the term \(\gamma _0\) is different from the baseline value of SHAP.

To better examine this aspect, we focus on a set of instances and analyze the explanations provided by the two models. Specifically, we consider the instance resulting in a high AE for both models (ID #2496) and all the instances of the test set that activate the same rule (namely, \(R_{223} \) ) of the TSK-FRBS. In this way, we isolate four instances (ID #827, ID #2266, ID #2496, ID #5146) which are inevitably close to each other in the input space.

The local interpretability of the TSK-FRBS is straightforward: predictions are obtained by applying the following rule:

In the case of MLP-NN, the prediction is explained through the Shapley values: Fig.  13 shows the contributions for the MLP-NN considering the four instances of the test set that activate rule \(R_{233} \) of the TSK-FRBS system.

It is evident that the Shapley values for the MLP-NN vary greatly even though the instances are fairly close in the input space: as an example, age has a negative contribution for ID #827 and a positive one for ID #5146. For this reason, it is equally evident that a correspondence cannot be found between the explanations offered by SHAP for the MLP-NN and the interpretation of the linear model of the TSK-FRBS. For example, SHAP always assigns a positive contribution to DFA , while the relevant coefficient is negative for TSK-FRBS.

As noted above, the divergence in explanations between TSK-FRBS and MLP-NN in the IID scenario does not correspond to a divergence in output values. We have verified that the predicted outputs are similar (and indeed the reported AE values are similar): different models, which achieve similar results, lead to different explanations from a local point of view. This is not to be considered odd: our analysis entails different feature importance methods (inherent and post-hoc) and different models (TSK-FRBS and MLP-NN, respectively). Actually, it has been empirically shown that the feature importance score may suffer from numerical instability (when model, instance and attribution method are the same), solution diversity (if different models are considered, but on the same instance with the same attribution methods), or disagreement problem (if different attribution methods are considered, but on the same instance and the same model) [ 83 ]. These scenarios are due to the so-called Rashomon effect [ 84 ], whereby for a given dataset there may exist many models with equally good performance but with different solution strategies.

Consistency of Explanations

As mentioned in Section 2.3 , the property of consistency in the FL setting, introduced in [ 60 ], is met if different participants receive the same explanation of an output obtained with the federated model given the same data instance.

Evidently, for all the different operative scenarios discussed in this paper, the explanations for the federated TSK-FRBS are consistent: any local explanation obtained in a given hospital depends only on the input instance and on the activated rule of the federated model.

Conversely, the approach proposed in [ 21 ] and adopted in this work as post-hoc technique for the MLP-NN explainability does ensure the consistency of the explanations only in the situation where the test instances are shareable to all the clients. The local explanation obtained in a given hospital, in fact, depends not only on the input instance and the federated MLP-NN, but also on the background dataset used for estimating the Shapley values. Since each hospital has its own private dataset, the Shapley values for the same input instance may differ, in general, from one hospital to another. The Federated SHAP approach allows obtaining an explanation for each test instance by averaging the local explanations from different hospitals. On one hand, this ensures that a unique and unambiguous explanation is obtained. On the other hand, this requires that any test instance is shared with other hospitals at inference time, which may be problematic due to privacy and/or latency issues.

Shapley values for the MLP-NN for instance ID #2496 for each hospital. IID Scenario

Shapley values for the MLP-NN for instance ID #2496 for each hospital. NIID-FQ scenario

In the following, we quantitatively assess the misalignment of client-side explanations obtained with SHAP for the MLP-NN, before applying the averaging operation that characterizes Federated SHAP. We consider an input test instance (ID# 2496, discussed also in the previous examples) and suppose that it is available at every hospital. Specifically, we evaluate how the prediction for such instance would be explained on different clients, in case that sharing Shapley values for averaging is precluded for privacy reason.

Figure  14 reports the Shapley values for each hospital and each feature in the IID scenario.

The barplot suggests that the explanations are consistent in the IID scenario, albeit showing some slight variability, which is reasonable since the background datasets are identically distributed. Indeed, explanations are in line with the average pattern reported in Fig.  14 c.

The variability among client-side explanations turns out to be substantial in non-i.i.d. scenarios. Figure  15 shows the Shapley values for the same instance for each hospital and each feature when considering the NIID-FQ scenario. We recall that such scenario entails both a quantity skew and a feature skew on the age feature. Furthermore, it is worth mentioning that the feature selection process is part of the FL pipeline. As a consequence, the set of selected features depends on the data distribution scenario: this explains the presence of different features compared to the IID case, with Jitter(Abs) replacing HNR .

Figure  15 suggests that the misalignment of explanations is severe, especially for the contribution assigned to the age feature by different hospitals. The relevant Shapley value goes from negative in hospitals with younger patients to positive in hospitals with older patients. Probing the global model with data from heterogeneous distributions results in a difference also in the importance assigned to the DFA feature. In summary, the same instance, analyzed with the same model, is explained in very different ways on different hospitals. Thus, the property of consistency among explanations is not achieved.

The consistency of SHAP explanations in the FL setting can be achieved by avoiding the use of private training data as background: Chen et al. [ 63 ], for example, propose to use synthetic background datasets generated sampling from a Gaussian distribution whose parameters are estimated on the server side based on the contributions of all participants. However, such explanations may be different from those obtained using actual training data. Ensuring both consistency and accuracy of explanations, intended as agreement with the centralized case, is one of the interesting future developments of this work.

Conclusions

In this paper, we have addressed the problem of developing a trustworthy AI system for a healthcare application, with specific focus on a Parkinson’s disease (PD) progression prediction task. For this purpose, we designed two approaches that simultaneously meet the requirements of data privacy preservation and explainability, which are usually deemed crucial for enabling trustworthiness. The first approach adopts a Takagi-Sugeno-Kang Fuzzy Rule-Based system (TSK-FRBS), which is interpretable by-design. TSK-FRBSs make use of fuzzy sets as information granules, thus guaranteeing high semantic interpretability. The second approach employs a Multi Layer Perceptron Neural Network (MLP-NN): as a “black-box” model, it requires the adoption of a post-hoc technique for explainability purposes. In this paper, we have adopted SHAP, which is considered as a state of art feature importance explainability method.

For both approaches, the federated learning (FL) paradigm has been exploited as it inherently enables privacy preservation during global model training procedures in decentralized settings. In detail, we devised an experimental setting assuming that sensitive data originate from ten hospitals and cannot be shared for privacy reasons. In order to cover several real-world situations, four (one i.i.d. and three non-i.i.d.) scenarios with different degrees of heterogeneity are simulated.

The critical analysis of the two approaches has concerned the following aspects: (i) the accuracy of the models, in terms of Root Mean Squared Error (RMSE) and Pearson correlation coefficient \(r,\,{ dependingonthelearningsetting}({ federatedlearning},\,{ locallearning},\,{ centralizedlearning}){ andthefourdatadistributionscenarios},\,{ and}({ ii}){ theexplainabilityofthemodelsatglobalandlocallevels}.{ Thekeyfindingscanbesummarizedasfollows}.{ Fromtheperspectiveofperformancemetrics},\,{ resultshighlightthatthefederatedmodelsareabletooutperformtheoneslearnedbyusingonlylocaldata},\,{ bothintermsofRMSEand}r\) values, hence highlighting the benefits of the federation. This is particularly evident in the non-i.i.d. settings. Also, results suggest that federated TSK-FRBS and federated MLP-NN achieve comparable performance, within the context of the considered case study.

As regards explainability, we have presented the results of the post-hoc explainability of MLP-NN and of the by-design interpretability of TSK-FRBS, also providing a comparative analysis of the two approaches. It turns out that the two approaches can lead to different local explanations, even if the underlying models achieve similar results in terms of regression metrics. A first major difference between the two approaches lies in their nature: the TSK-FRBS model provides insights about “how” an outcome is obtained, whereas the post-hoc method provides insight about “why” an outcome is provided. Consequently, a qualitative comparison is more reasonable than a quantitative one. We have to take into account, however, that if a model is able to provide a glimpse on how an outcome has been obtained, implicitly it is also making manifest why that outcome has been computed from the inputs.

More precisely, the TSK-FRBS is a collection of linguistic if-then rules: its global interpretability is usually assessed in terms of the number of rules and/or parameters: a lower number of rules corresponds to systems considered more interpretable. An equivalent global picture is not immediately available in the case of MLP-NN: the adoption of the SHAP method enables local explainability (i.e., it explains individual predictions), and a global explainability indication can be obtained in terms of feature importance, by aggregating the individual explanations over a set of data. The comparison of the feature importance obtained for the MLP-NN (as the average of the absolute Shapley values) and for the TSK-FRBS (as the average of the absolute values of the coefficients of the linear models in the rule base) in the IID setting reveals a fair agreement: both methods identify age as the most important feature and test_time as the least important one.

Local explanations of the two approaches convey inherently different messages. For the MLP-NN, the Shapley values represent the importance of each feature to each prediction with respect to a baseline value, computed as the average output value. For the TSK-FRBS, the predicted output depends, in our setting, just on the mostly activated rule: the antecedent part isolates a region of the input space and is expressed using linguistic terms associated with fuzzy sets (information granules), whereas its consequent part defines a local linear model whose coefficients indicate how the input features contribute to form the output. Focusing on a specific region of the input space, corresponding to the one isolated by the antecedent of a rule in the TSK-FRBS, we have evaluated the explanations provided by the two models for the instances located within that region. As expected, the explanations obtained with SHAP for the MLP-NN are generally different for the different instances and do not match what is expressed by the relevant rule of the TSK-FRBS. In essence, different models, which produce similar outputs and achieve similar overall results, lead to different explanations from a local point of view.

Finally, it is worth underlining that the property of consistency holds for the federated TSK-FRBS: for the same input data, different participants get the same explanation from the common federated model. In the case of the SHAP post-hoc technique, this is not as straightforward: if participants use their local training data as background dataset for the estimation of the Shapley values, the feature importance scores are necessarily different; if the federated SHAP strategy of averaging locally computed Shapley values is adopted, a single value is obtained. This ensures consistency of explanation but requires that the privacy preservation constraint is relaxed at test time. How to achieve consistency of post-hoc explanation and simultaneously to ensure privacy preservation represents an interesting future development of this work.

Overall, the Fed-XAI approach offers remarkable potential and can have important practical implications in high-stake applications such as healthcare: in this domain, in fact, data centralization for ML model training is not only difficult to implement from a technical perspective, but it also raises ethical concerns related to privacy and is subject to limitations imposed by regulatory policies. On one hand, FL removes the need for data centralization while still allowing ML models training to benefit from ample and diverse data, which is crucial to address urgent challenges such as health disparities, under-served populations, and rare diseases [ 85 ]. On the other hand, the adoption of XAI tools endow the AI system with the capability of explaining its decisions, which is paramount in these kinds of applications.

We have discussed strengths and weaknesses of the two federated approaches for learning XAI models. In the future, we aim to enrich the comparative analysis with an additional level of assessment by collecting feedbacks on explainability from human experts (e.g., physicians) and by considering other case studies.

Data Availability

The dataset is publicly available at https://archive.ics.uci.edu/dataset/189/parkinsons+telemonitoring .

https://artificialintelligenceact.eu/the-act/ , visited May 2024.

High level expert group on AI. Ethics guidelines for trustworthy AI, Technical Report. European Commission. https://ec.europa.eu/digital-single-market/en/news/ethics-guidelines-trustworthy-ai .

: GDPR. Available from: https://gdpr-info.eu/recitals/no-71/ [cited 06.10.2022].

Barredo Arrieta A, Díaz-Rodríguez N, Del Ser J, Bennetot A, Tabik S, Barbado A, et al. Explainable Artificial Intelligence (XAI): concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion. 2020;58:82–115. https://doi.org/10.1016/j.inffus.2019.12.012 .

Article   Google Scholar  

Guidotti R, Monreale A, Ruggieri S, Turini F, Giannotti F, Pedreschi D. A survey of methods for explaining black box models. ACM Comput Surv. 2018. https://doi.org/10.1145/3236009 .

Guidotti R, Monreale A, Pedreschi D, Giannotti F. In: Sayed-Mouchaweh M, editor. Principles of explainable artificial intelligence. Cham: Springer International Publishing; 2021. p. 9–31.

Ali S, Abuhmed T, El-Sappagh S, Muhammad K, Alonso-Moral JM, Confalonieri R, et al. Explainable Artificial Intelligence (XAI): what we know and what is left to attain Trustworthy Artificial Intelligence. Information Fusion. 2023;99:101805. https://doi.org/10.1016/j.inffus.2023.101805 .

McMahan B, Moore E, Ramage D, Hampson S, Arcas BAy. Communication-efficient learning of deep networks from decentralized data. In: Singh A, Zhu J, editors. Proceedings of the 20th international conference on artificial intelligence and statistics. vol. 54 of Proceedings of Machine Learning Research. PMLR; 2017. p. 1273–1282. Available from: https://proceedings.mlr.press/v54/mcmahan17a.html .

Yang Q, Liu Y, Chen T, Tong Y. Federated machine learning: concept and applications. ACM Trans Intell Syst Technol. 2019. https://doi.org/10.1145/3298981 .

Bodria F, Giannotti F, Guidotti R, Naretto F, Pedreschi D, Rinzivillo S. Benchmarking and survey of explanation methods for black box models. Data Mining and Knowledge Discovery. 2023;37(5):1719–78. https://doi.org/10.1007/s10618-023-00933-9 .

Article   MathSciNet   Google Scholar  

Chaddad A, Lu Q, Li J, Katib Y, Kateb R, Tanougast C, et al. Explainable, domain-adaptive, and federated artificial intelligence in medicine. IEEE/CAA Journal of Automatica Sinica. 2023;10(4):859–76. https://doi.org/10.1109/JAS.2023.123123 .

Corcuera Bárcena JL, Daole M, Ducange P, Marcelloni F, Renda A, Ruffini F, et al. Fed-XAI: Federated Learning of Explainable Artificial Intelligence Models. In: XAI.it: 3rd Italian workshop on explainable artificial intelligence, co-located with AI*IA; 2022. Available from: https://ceur-ws.org/Vol-3277/paper8.pdf .

López-Blanco R, Alonso RS, González-Arrieta A, Chamoso P, Prieto J. Federated Learning of Explainable Artificial Intelligence (FED-XAI): a review. In: Ossowski S, Sitek P, Analide C, Marreiros G, Chamoso P, Rodríguez S, editors. Distributed computing and artificial intelligence, 20th international conference. Cham: Springer Nature Switzerland; 2023. p. 318–26.

Google Scholar  

Corcuera Bárcena JL, Ducange P, Ercolani A, Marcelloni F, Renda A. An approach to federated learning of explainable fuzzy regression models. In: IEEE International conference on fuzzy systems (FUZZ-IEEE); 2022. p. 1–8.

Daneault JF, Carignan B, Sadikot AF, Duval C. Are quantitative and clinical measures of bradykinesia related in advanced Parkinson’s disease? [Article]. Journal of Neuroscience Methods. 2013;219(2):220–3. https://doi.org/10.1016/j.jneumeth.2013.08.009 .

Corcuera Bárcena JL, Ducange P, Marcelloni F, Renda A, Ruffini F. Federated learning of explainable artificial intelligence models for predicting Parkinson’s disease progression. In: Longo L, editor. Explainable artificial intelligence. Cham: Springer Nature Switzerland; 2023. p. 630–48.

Chapter   Google Scholar  

Takagi T, Sugeno M. Fuzzy identification of systems and its applications to modeling and control. IEEE Transactions on Systems, Man, and Cybernetics. 1985;SMC–15(1):116–32. https://doi.org/10.1109/TSMC.1985.6313399 .

Fernandez A, Herrera F, Cordon O, Jose del Jesus M, Marcelloni F. Evolutionary fuzzy systems for explainable artificial intelligence: why, when, what for, and where to? Comp Intell Mag. 2019;14(1):69–81. https://doi.org/10.1109/MCI.2018.2881645 .

Zadeh LA. Toward a theory of fuzzy information granulation and its centrality in human reasoning and fuzzy logic. Fuzzy Sets and Systems. 1997;90(2):111–27. https://doi.org/10.1016/S0165-0114(97)00077-8 . ( Fuzzy Sets: Where Do We Stand? Where Do We Go? ).

Yao JT, Vasilakos AV, Pedrycz W. Granular computing: perspectives and challenges. IEEE Transactions on Cybernetics. 2013;43(6):1977–89. https://doi.org/10.1109/TSMCC.2012.2236648 .

Lundberg SM, Lee SI. A unified approach to interpreting model predictions. In: Guyon I, Luxburg UV, Bengio S, Wallach H, Fergus R, Vishwanathan S, et al., editors. Advances in Neural Information Processing Systems. vol. 30. Curran Associates, Inc.; 2017. Available from: https://proceedings.neurips.cc/paper_files/paper/2017/file/8a20a8621978632d76c43dfd28b67767-Paper.pdf .

Corbucci L, Guidotti R, Monreale A. Explaining black-boxes in federated learning. In: Longo L, editor. Explainable artificial intelligence. Cham: Springer Nature Switzerland; 2023. p. 151–63.

Li H, Li C, Wang J, Yang A, Ma Z, Zhang Z, et al. Review on security of federated learning and its application in healthcare. Future Generation Computer Systems. 2023;144:271–90. https://doi.org/10.1016/j.future.2023.02.021 .

Hwang H, Yang S, Kim D, Dua R, Kim JY, Yang E, et al. Towards the practical utility of federated learning in the medical domain. In: Mortazavi BJ, Sarker T, Beam A, Ho JC, editors. Proceedings of the conference on health, inference, and learning. vol. 209 of Proceedings of Machine Learning Research. PMLR; 2023. p. 163–181. Available from: https://proceedings.mlr.press/v209/hwang23a.html .

Rahman A, Hossain MS, Muhammad G, Kundu D, Debnath T, Rahman M, et al. Federated learning-based AI approaches in smart healthcare: concepts, taxonomies, challenges and open issues. Cluster computing. 2023;26(4):2271–311. https://doi.org/10.1007/s10586-022-03658-4 .

Sohan MF, Basalamah A. A systematic review on federated learning in medical image analysis. IEEE Access. 2023;11:28628–44. https://doi.org/10.1109/ACCESS.2023.3260027 .

De Falco I, Della Cioppa A, Koutny T, Scafuri U, Tarantino E. Model-free-communication federated learning: framework and application to precision medicine. Biomedical Signal Processing and Control. 2024;87:105416. https://doi.org/10.1016/j.bspc.2023.105416 .

Nguyen DC, Pham QV, Pathirana PN, Ding M, Seneviratne A, Lin Z, et al. Federated learning for smart healthcare: a survey [article]. ACM Computing Surveys. 2022. https://doi.org/10.1145/3501296 . ( Cited by: 209; All Open Access, Green Open Access ).

Ali M, Naeem F, Tariq M, Kaddoum G. Federated learning for privacy preservation in smart healthcare systems: a comprehensive survey [article]. IEEE Journal of Biomedical and Health Informatics. 2023;27(2):778–89. https://doi.org/10.1109/JBHI.2022.3181823 . ( Cited by: 36; All Open Access, Green Open Access ).

Hassija V, Chamola V, Mahapatra A, Singal A, Goel D, Huang K, et al. Interpreting black-box models: a review on explainable artificial intelligence. Cognitive Computation. 2023. https://doi.org/10.1007/s12559-023-10179-8 .

Patrício C, JaC Neves, Teixeira LF. Explainable deep learning methods in medical image classification: a survey. ACM Comput Surv. 2023. https://doi.org/10.1145/3625287 .

Uddin MZ, Dysthe KK, Følstad A, Brandtzaeg PB. Deep learning for prediction of depressive symptoms in a large textual dataset. Neural Computing and Applications. 2022;34(1):721–44. https://doi.org/10.1007/s00521-021-06426-4 .

El-Sappagh S, Alonso JM, Islam SMR, Sultan AM, Kwak KS. A multilayer multimodal detection and prediction model based on explainable artificial intelligence for Alzheimer’s disease. Sci Rep. 2021;11(1):2660.

Dong N, Voiculescu I. Federated contrastive learning for decentralized unlabeled medical images. In: Medical Image Computing and Computer Assisted Intervention - MICCAI 2021: 24th International Conference, Strasbourg, France, September 27-October 1, 2021, Proceedings, Part III. Berlin, Heidelberg: Springer-Verlag; 2021. p. 378–87. https://doi.org/10.1007/978-3-030-87199-4_36 .

Vaid A, Jaladanki SK, Xu J, Teng S, Kumar A, Lee S, et al. Federated learning of electronic health records to improve mortality prediction in hospitalized patients with COVID-19: machine learning approach. JMIR Med Inform. 2021;9(1):e24207.

Bounsall K, Milne-Ives M, Hall A, Carroll C, Meinert E. Artificial intelligence applications for assessment, monitoring, and management of parkinson disease symptoms: protocol for a systematic review. JMIR Res Protoc. 2023;12:e46581. https://doi.org/10.2196/46581 .

El-Sappagh S, Alonso-Moral JM, Abuhmed T, Ali F, Bugarín-Diz A. Trustworthy artificial intelligence in Alzheimer’s disease: state of the art, opportunities, and challenges. Artif Intell Rev. 2023;56(10):11149–296. https://doi.org/10.1007/s10462-023-10415-5 .

Gómez-Vilda P, Rodellar-Biarge V, Nieto-Lluis V, Muñoz-Mulas C, Mazaira-Fernández LM, Martínez-Olalla R, et al. Characterizing neurological disease from voice quality biomechanical analysis. Cognitive Computation. 2013;5(4):399–425. https://doi.org/10.1007/s12559-013-9207-2 .

Arias-Vergara T, Vásquez-Correa JC, Orozco-Arroyave JR. Parkinson’s disease and aging: analysis of their effect in phonation and articulation of speech. Cognitive Computation. 2017;9(6):731–48. https://doi.org/10.1007/s12559-017-9497-x .

Magesh P, Myloth R, Tom R. An explainable machine learning model for early detection of Parkinson’s disease using lime on DaTSCAN imagery. Computers in Biology and Medicine. 2020;11(126):104041. https://doi.org/10.1016/j.compbiomed.2020.104041 .

Ribeiro MT, Singh S, Guestrin C. “Why should i trust you?”: explaining the predictions of any classifier. In: Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining. KDD ’16. New York, USA: Association for Computing Machinery; 2016. p. 1135-1144.

Junaid M, Ali S, Eid F, El-Sappagh S, Abuhmed T. Explainable machine learning models based on multimodal time-series data for the early detection of Parkinson’s disease. Computer Methods and Programs in Biomedicine. 2023;234:107495. https://doi.org/10.1016/j.cmpb.2023.107495 .

Jorge J, Barros PH, Yokoyama R, Guidoni D, Ramos HS, Fonseca N, et al. Applying federated learning in the detection of freezing of gait in Parkinson’s disease. In: IEEE/ACM 15th International conference on Utility and Cloud Computing (UCC); 2022. p. 195–200.

Sarlas A, Kalafatelis A, Alexandridis G, Kourtis MA, Trakadas P. Exploring federated learning for speech-based Parkinson’s disease detection. In: Proceedings of the 18th international conference on availability, reliability and security. ARES ’23. New York, USA: Association for Computing Machinery; 2023.

Dipro SH, Islam M, Al Nahian A, Sharmita Azad M, Chakrabarty A, Reza T. A federated learning based privacy preserving approach for detecting Parkinson’s disease using deep learning. In: 2022 25th International Conference on Computer and Information Technology (ICCIT); 2022. p. 139–144.

Grover S, Bhartia S, Akshama, Yadav A, R SK. Predicting severity of Parkinson’s disease using deep learning. Procedia Computer Science. Procedia Computer Science. 2018;132:1788–94. https://doi.org/10.1016/j.procs.2018.05.154 . ( International Conference on Computational Intelligence and Data Science ).

Gunduz H. Deep sets learning-based Parkinson’s disease classification using vocal feature. IEEE Access. 2019;7:115540–51. https://doi.org/10.1109/ACCESS.2019.2936564 .

Nilashi M, Ibrahim O, Samad S, Ahmadi H, Shahmoradi L, Akbari E. An analytical method for measuring the Parkinson’s disease progression: a case on a Parkinson’s telemonitoring dataset. Measurement. 2019;136:545–57. https://doi.org/10.1016/j.measurement.2019.01.014 .

Shahid AH, Singh MP. A deep learning approach for prediction of Parkinson’s disease progression. Biomedical Engineering Letters. 2020;10:227–39. https://doi.org/10.1007/s13534-020-00156-7 .

Xue Z, Zhang T, Lin L. Progress prediction of Parkinson’s disease based on graph wavelet transform and attention weighted random forest. Expert Systems with Applications. 2022;203: 117483. https://doi.org/10.1016/j.eswa.2022.117483 .

Chen P, Du X, Lu Z, Wu J, Hung PCK. EVFL: an explainable vertical federated learning for data-oriented Artificial Intelligence systems. Journal of Systems Architecture. 2022;126:102474. https://doi.org/10.1016/j.sysarc.2022.102474 .

Fiosina J. Explainable federated learning for taxi travel time prediction. vol. 2021-April; 2021. p. 670 - 677. Available from: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85121125924&partnerID=40 &md5=9a076bc80768b97eee10c0883e2557e7 .

Fiosina J. Interpretable privacy-preserving collaborative deep learning for taxi trip duration forecasting. In: International conference on vehicle technology and intelligent transport systems, international conference on smart cities and green ICT systems. Springer; 2022. p. 392–411.

Wang G. Interpret federated learning with Shapley values. arXiv preprint arXiv:1905.04519 . 2019;.

Sidhpura J, Shah P, Veerkhare R, Godbole A. FedSpam: privacy preserving SMS spam prediction. In: Tanveer M, Agarwal S, Ozawa S, Ekbal A, Jatowt A, editors. Neural information processing. Singapore: Springer Nature Singapore; 2023. p. 52–63.

Ben Saad S, Brik B, Ksentini A. A trust and explainable federated deep learning framework in zero touch B5G networks. In: GLOBECOM - IEEE Global Communications Conference; 2022. p. 1037–1042.

Ludwig H, Baracaldo N, Thomas G, Zhou Y, Anwar A, Rajamoni S, et al.: IBM federated learning: an enterprise framework white paper V0.1. arXiv. Available from: https://arxiv.org/abs/2007.10987 .

Wilbik A, Grefen P. Towards a federated fuzzy learning system. In: IEEE International conference on fuzzy systems (FUZZ-IEEE); 2021. p. 1–6.

Wu Y, Cai S, Xiao X, Chen G, Ooi BC. Privacy preserving vertical federated learning for tree-based models. Proc VLDB Endow. 2020;13(12):2090–103. https://doi.org/10.14778/3407790.3407811 .

Zhu X, Wang D, Pedrycz W, Li Z. Horizontal federated learning of Takagi-Sugeno fuzzy rule-based models. IEEE Transactions on Fuzzy Systems. 2022;30(9):3537–47. https://doi.org/10.1109/TFUZZ.2021.3118733 .

Bogdanova A, Imakura A, Sakurai T. DC-SHAP method for consistent explainability in privacy-preserving distributed machine learning. Human-Centric Intelligent Systems. 2023;3(3):197–210. https://doi.org/10.1007/s44230-023-00032-4 .

Zheng S, Cao Y, Yoshikawa M. Secure Shapley value for cross-silo federated learning. Proc VLDB Endow. 2023, 16(7), pp. 1657–1670. https://doi.org/10.14778/3587136.3587141

Janzing D, Minorics L, Blöbaum P. Feature relevance quantification in explainable AI: a causal problem. In: International Conference on artificial intelligence and statistics. PMLR; 2020. p. 2907–2916. Available from: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85122610367&partnerID=40 &md5=32e4054213ccb02f05bfba648c864fbb .

Chen Y, Yang X, He Y, Miao C, Chan P. FedDBM: federated digital biomarker for detecting Parkinson’s disease progress. In: IEEE International Conference on Multimedia and Expo (ICME); 2023. p. 678–683.

Polato M, Esposito R, Aldinucci M. Boosting the federation: cross-silo federated learning without gradient descent. In: International Joint Conference on Neural Networks (IJCNN); 2022. p. 1–10.

Alonso Moral JM, Castiello C, Magdalena L, Mencar C. In: Designing interpretable fuzzy systems. Cham: Springer International Publishing; 2021. p. 119–168. Available from: https://doi.org/10.1007/978-3-030-71098-9_5 .

Gacto MJ, Alcalá R, Herrera F. Interpretability of linguistic fuzzy rule-based systems: an overview of interpretability measures. Information Sciences. 2011;181(20):4340–60. https://doi.org/10.1016/j.ins.2011.02.021 . ( Special Issue on Interpretable Fuzzy Systems ).

Zhu H, Xu J, Liu S, Jin Y. Federated learning on non-IID data: a survey. Neurocomputing. 2021;465:371–90. https://doi.org/10.1016/j.neucom.2021.07.098 .

Morafah M, Wang W, Lin B. A practical recipe for federated learning under statistical heterogeneity experimental design. IEEE Transactions on Artificial Intelligence. 2023;p. 1–14. https://doi.org/10.1109/TAI.2023.3297090 .

Shapley LS. In: Kuhn HW, Tucker AW, editors. 17. A value for n-person games. Princeton University Press: Princeton; 1953. p. 307–18.

Molnar C.: Interpretable machine learning: a guide for making black box models explainable. Available from: https://christophm.github.io/interpretable-ml-book .

Tsanas A, Little MA, McSharry PE, Ramig LO. Accurate telemonitoring of Parkinson’s disease progression by noninvasive speech tests. IEEE Transactions on Biomedical Engineering. 2010;57(4):884–93. https://doi.org/10.1109/TBME.2009.2036000 .

Kairouz P, McMahan HB, Avent B, Bellet A, Bennis M, Bhagoji AN, et al.: Advances and open problems in federated learning. Available from: http://dx.doi.org/10.1561/2200000083 .

Bakopoulou E, Tillman B, Markopoulou A. FedPacket: a federated learning approach to mobile packet classification. IEEE Transactions on Mobile Computing. 2022;21(10):3609–28. https://doi.org/10.1109/TMC.2021.3058627 .

Corcuera Bárcena JL, Ducange P, Marcelloni F, Nardini G, Noferi A, Renda A, et al. Enabling federated learning of explainable AI models within beyond-5G/6G networks. Computer Communications. 2023;210:356–75. https://doi.org/10.1016/j.comcom.2023.07.039 .

Cover TM. Elements of information theory. John Wiley & Sons; 1999.

Cózar J, Ossa Ldl, Gámez JA. TSK-0 fuzzy rule-based systems for high-dimensional problems using the apriori principle for rule generation. In: International Conference on Rough Sets and Current Trends in Computing. Springer; 2014. p. 270–279.

Li Q, Wen Z, Wu Z, Hu S, Wang N, Li Y, et al. A survey on federated learning systems: vision, hype and reality for data privacy and protection. IEEE Transactions on Knowledge and Data Engineering. 2023;35(4):3347–66. https://doi.org/10.1109/TKDE.2021.3124599 .

Alonso JM, Magdalena L, Guillaume S. HILK: a new methodology for designing highly interpretable linguistic knowledge bases using the fuzzy logic formalism. International Journal of Intelligent Systems. 2008;23(7):761–94. https://doi.org/10.1002/int.20288 .

Miller GA. The magical number seven, plus or minus two: some limits on our capacity for processing information. Psychological Review. 1956;63(2):81–97. https://doi.org/10.1037/h0043158 .

Wilcoxon F. Individual comparisons by ranking methods. In: Breakthroughs in statistics. Springer; 1992. p. 196–202.

Fuchs C, Spolaor S, Nobile MS, Kaymak U. pyFUME: a Python package for fuzzy model estimation. In: IEEE International Conference on Fuzzy Systems (FUZZ-IEEE); 2020. p. 1–8.

Obeso JA, Rodriguez-Oroz MC, Goetz CG, Marin C, Kordower JH, Rodriguez M, et al. Missing pieces in the Parkinson’s disease puzzle [Review]. Nature Medicine. 2010;16(6):653–61. https://doi.org/10.1038/nm.2165 .

Müller S, Toborek V, Beckh K, Jakobs M, Bauckhage C, Welke P. An empirical evaluation of the Rashomon effect in explainable machine learning. In: Koutra D, Plant C, Gomez Rodriguez M, Baralis E, Bonchi F, editors. Machine learning and knowledge discovery in databases: Research Track. Cham: Springer Nature Switzerland; 2023. p. 462–78.

Breiman L. Statistical modeling: the two cultures (with comments and a rejoinder by the author). Statistical Science. 2001;16(3):199–231. https://doi.org/10.1214/ss/1009213726

Pati S, Baid U, Edwards B, Sheller M, Wang SH, Reina GA, et al. Federated learning enables big data for rare cancer boundary detection. Nature Communications. 2022. https://doi.org/10.1038/s41467-022-33407-5 .

Download references

Open access funding provided by Università di Pisa within the CRUI-CARE Agreement. This work has been partly funded by the PON 2014–2021 “Research and Innovation,” DM MUR 1062/2021, Project title: “Progettazione e sperimentazione di algoritmi di federated learning per data stream mining,” PNRR - M4C2 - Investimento 1.3, Partenariato Esteso PE00000013 - “FAIR - Future Artificial Intelligence Research” - Spoke 1 “Human-centered AI” and the PNRR “Tuscany Health Ecosystem” (THE) (Ecosistemi dell’Innovazione) - Spoke 6 - Precision Medicine & Personalized Healthcare (CUP I53C22000780001) under the NextGeneration EU programme, and by the Italian Ministry of University and Research (MUR) in the framework of the FoReLab and CrossLab projects (Departments of Excellence).

Author information

Authors and affiliations.

Department of Information Engineering, University of Pisa, Largo Lucio Lazzarino 1, Pisa, 56122, Italy

Pietro Ducange, Francesco Marcelloni, Alessandro Renda & Fabrizio Ruffini

You can also search for this author in PubMed   Google Scholar

Contributions

All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Alessandro Renda and Fabrizio Ruffini. The first draft of the manuscript was written by Francesco Marcelloni, Alessandro Renda, and Fabrizio Ruffini, and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Alessandro Renda or Fabrizio Ruffini .

Ethics declarations

Conflict of interest.

The authors declare no competing interests.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Ducange, P., Marcelloni, F., Renda, A. et al. Federated Learning of XAI Models in Healthcare: A Case Study on Parkinson’s Disease. Cogn Comput (2024). https://doi.org/10.1007/s12559-024-10332-x

Download citation

Received : 15 December 2023

Accepted : 23 July 2024

Published : 28 August 2024

DOI : https://doi.org/10.1007/s12559-024-10332-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Federated learning
• Explainable artificial intelligence
• Granular fuzzy models
• Parkinson’s disease

Advertisement

• Find a journal
• Publish with us
• Track your research